Performance enhancement of Building-Integrated Concentrator Photovoltaic system using Phase Change Materials

Abstract Building-Integrated Concentrated Photovoltaic (BICPV) systems integrate easily into built environments, replacing building material, providing benefits of generating electricity at the point of use, allowing light efficacy within the building envelope and providing thermal management. This paper presents a novel experimental evaluation of phase change materials (PCM) to enhance performance of low-concentration BICPV system via thermal regulation. Previous studies have primarily focussed on temporal and spatial studies of PCM temperature within the BIPV systems but the current work also discusses the effect of PCM on electrical parameters of the BICPV systems. Due to the inadequacy of the earlier reported model, a new analytical model is proposed and implemented with the in-house controlled experiments. Paraffin wax based RT42 was used within an in-house designed and fabricated PCM containment. An indoor experiment was performed using highly collimated continuous light source at 1000 W m −2 . Results show an increase in relative electrical efficiency by 7.7% with PCM incorporation. An average reduction in module centre temperature by 3.8 °C was recorded in the BICPV–PCM integrated system as compared to the naturally ventilated system without PCM. Studies showed that PCM effectiveness varies with irradiance; an increase in relative electrical efficiency by 1.15% at 500 W m −2 , 4.20% at 750 W m −2 and 6.80% at 1200 W m −2 was observed.


List of Figures
Heat transfer coefficient of the un-finned surface (Wm -2 K -1 ) Heat transfer coefficient at front surface (Wm -2 K -1 ) Heat transfer coefficient at rear surface (Wm -2 K -1 ) This chapter starts with an overview on PV, CPV and BICPV and reviews the state of the art of these technologies along with the associated technical parameters in the first part.
PV is further categorised into BAPV (building-attached photovoltaic) and BIPV (building-integrated photovoltaic). From here, it follows on with an introduction to the different types of solar concentrators; a feature that differentiates a traditional PV from a CPV. The advantage and limitations of these technologies, areas of existing as well as proposed applications, current challenges and some available solutions pertaining to each of these areas are discussed. In latter sections following the background information, the important research questions related to challenges with BICPV thermal regulation, the rationale for pursuing this research, the key aims and the employed research methodology are summarised. Finally, in the last part, the salient features of this research work based on low concentration BICPV, the distinguishable in-scope elements and the thesis organisation are elucidated.

Background
Solar energy, with its tremendous availability of 120 petajoules per second on the Earth shows immense potential, having annual net energy production increase by an average 8.3 % and a projected generation of 859 billion kWh, which was accountable for a 15 % of the 5.9 trillion kWh of new renewable generation by 2014 [1]. Over the last one and a half decades, PV markets have grown at a rate of 42 % compounded annually  out of which Si-based PV accounted for about 93 % (with 68 % contribution was from multi-crystalline) and about 7 % from thin film technologies to the total annual production in 2015 [2]. Owing to increasing efficiencies, the Si usage in cells in the last decade have reduced from around 16 g/Wp to 6 g/Wp. [2]. Fig. 1 shows the global PV power production capacity (in GW) and annual additions since 2005, when solar PV contributed to 40 GWp growing to a 227 GWp in 2015, an almost six-fold increase within a decade. Considering U.K. scenario (Fig. 2 [4]. The integration of PV into architectural design is both significant and innovative because buildings are responsible for a major proportion of total energy usage. The data for the primary energy consumption in the domestic sector buildings in the U.K. Fig. 3 shows that a major 66 % of the share goes towards space heating, followed by 17 % towards water heating. The use of PV with the buildings can act as a stand-alone or a grid-connected micro-renewable electricity energy source.
BIPV/BAPV have the potential to contribute to clean energy production at the site of use and unburden the load on the grid by decentralising electricity production.   [5].
A summary of the international BIPV research activities that have taken place over the last four decades clearly shows the exponential growth in the sector. In particular, the last decade has seen the evolution of new societies, task committees, international conferences and events, which signifies the presence of BIPV knowledge spreading globally (Fig. 4). innovative solutions based on less material requiring PV technologies [7]. They also offer a promising solution for electrification of rural areas cut off from main electricity grids, especially in developing countries. Besides, the recent government policies in many countries such as the European Union pushes towards energy-neutral buildings, which not only drive building integration of PV but also benefits by heat generation and daylight regulation by transmission of diffuse sunlight through transparent parts of the emerging PV systems [8]. However, there are policy-related and technical challenges to adopting BIPV, and more importantly, BICPV, on a larger scale, some of which are explored within this thesis and the potential solutions are suggested. Can be installed on-site for decentralised power production or grid-connected for centralised power production and distribution

PV: Building applications
Depending on the point of application, PV could be either retrofit to an existing and already constructed building (externally) or integrated with the building, replacing the building material (internally) wherein the PV panel forms an integral part of the building itself, hence the names building-applied or building-attached (BAPV) and buildingintegrated (BIPV) respectively. As a part of the building skin, PV can be used as a basic or functional construction element such as tile or cladding panel and as modular-unitized system combining BIPV modules with the building or electrical sub-constructions (prefabrication, plug & play etc.) [6]. BAPV systems are simpler and quicker in installation with lower installation costs but they increase the building load and due to repeated constructions, there is a waste of the building materials that isn't the case for BIPV systems as PV arrays are integrated directly into the building skeleton [11]. PV installers and commercial companies [12] believe that retrofitting or BAPV is "neither a good use of material resources nor does it achieve the best aesthetic, functional or cost BIPV, within the building envelopes, solve the primary purpose of generating electrical energy for use at the point of generation, however, the device itself can work as a replacement for construction elements such as roofs, facade claddings, windows, and static or dynamic shading elements [13] and are capable of delivering electricity at less than the cost of grid electricity to end users in certain peak demand niche markets [14].
BIPV are modular in nature, leading to reduced installation time with less regular maintenance needs and offer an alternative to the use of land for PV modules especially in urban areas where land availability is a challenge. In addition, the cost for free land needed for installation can be saved. The technical standards of BIPV are defined in BS EN 50583. BIPV system performance depends on (a) the solar radiation availability which in turn depends on the geographical location, climate and building orientation and (b) the collection ability of the device, which is determined by the PV efficiency, inclination etc. [15]. The integration of PV into buildings are based on many factors such as their design, purpose, building and safety regulations and aesthetics, which are schematically illustrated in Fig. 5. As building elements, BIPV can contribute in one or more of the following ways: by providing mechanical rigidity, structural integrity, weather protection, energy economy and thermal insulation, fire and noise protection, and function as a partition between the indoor and the outdoor surroundings [13].   [12]. For instance, a modern commercial design of using BIPV as a building element as shown in Fig. 6, where the completely recyclable, natural and fire resistant air incorporated rock-wool brings thermal insulation while the monocrystalline silicon solar cells protected by two layers of glass provide high structural rigidity forming a walkable roof for easy installation and maintenance [16]. In case of lightning, the protection is provided as all the elements are already connected and form the lightning grid, once connected to the ground, thus assuring the safety of occupants.  Although BIPV isn't a new concept, there are still many challenges related to the large scale adoption of this technology mostly due to the requirement of extensive planning and architectural challenges while further challenges include a lack of consistent markets, complexities in planning and construction process, a gap in available certified construction products, complicated installation methods requiring specialist know-how, lack of technical awareness in architects and planners and higher costs in addition to constraints related to locations such as solar radiation availability throughout the day and maintenance etc. [6].

CPV: an overview
Concentrator Photovoltaic (CPV) technology is based on the general principle of using optical elements (working as radiation collectors) such as mirrors or lenses, to focus solar irradiance on a solar cell area, which is much smaller, compared to a conventional PV panel. The benefits of using concentration are:  Increased solar radiation for photovoltaic conversion; higher electrical output  Reduced semiconductor material usage for the same power output; saving costs and environmental impacts caused during solar cell manufacturing

Components
A schematic of the components forming a CPV assembly is given in Fig. 8. The key elements forming a CPV system are reflective or refractive optics, solar cells soldered to form a structure for desired electrical output and an external casing to support the assembly. Additionally, thermal management could be achieved by using heat dissipation methods such as active or passive cooling. Tracking mechanism is required for accuracy in tracking the sun especially for high concentration systems based on more expensive multi-junction solar cells and it enables electricity production for prolonged hours. Each of these elements or components contributes individually and collectively to overall performance of CPV [17].

Concentration Parameters
The key parameters associated with concentrators are defined as:  Geometric concentration ratio (Cg) or the number of suns, is the ratio of collector (entry) aperture area and receiver (exit) aperture area of the concentrator and is given by Eqn.
Where: Ir is the averaged irradiance or radiant flux; and Ia is incident irradiance; (Wm -2 )  The optical efficiency of CPV is given by (Eqn. 4).
Where: IMax_Conc is the maximum current generated by the CPV; (A) and IMax is the maximum current generated by the PV (without the concentrator); (A).
Optical efficiency is an important parameter in determining the maximum limit of the concentration collection efficiency. Literature suggests that as it's possible to obtain over 90 % of optical efficiency which could contribute to the module efficiency reaching 45 % and full annual plant efficiency to increase from the current value of 12 % to 40 % [18]. The cost of concentration optics is generally lower than that of solar cells [19]. IEC 62108, a combination of two previous standards, IEEE 1513 and IEC 61215, are the primary standards for determining the electrical, mechanical, and thermal characterisation of CPV modules [20]. It is imperative to classify the concentrator type for deciding the feasibility of use and assessing the suitability for an application and the cooling mechanism to be employed alongside.

Classification
The classification of solar concentrators can be based on optical characteristics; concentration factor or concentration ratio, distribution of illumination, focal shape and optical standards [21]. CPV classification based on concentration mode, geometry, primary and secondary optics and applications are discussed below (Fig. 9).
LCPV can be easily building-integrated as they mostly do not require electromechanical tracking of the sun and optical gain is achieved under both direct and diffuse radiation [25]. LCPV with higher concentrations can work efficiently with single-axis tracking [22]. The higher end of MCPV (60 < Cg < 85) is not deemed suitable for building applications due to their size. They can be used for power plants and many MCPV designs differ from typical planar modules. The optical precision and tracker accuracy requirements are not extreme but there are challenges during cell mounting for good voltage standoff and thermal management [26]. For HCPV, however, very high precision two-axis tracking is mandatory, which is why HCPV can't be used for building integration. They also cause architectural and aesthetics issues. In some cases of HCPV, highly efficient but expensive solar cells are used, but concentration can still provide economic viability since the area of the cells can be up to about 100 times smaller than the light collection area [27].
Fresnel lenses, Cassegrain optics and point concentration reflector are generally used for HCPV applications while MCPV relies on either parabolic troughs or Fresnel optics (lens or mirror). LCPV uses either V-trough reflectors or compound parabolic concentrator (CPC) [28]. V-troughs direct light onto the receiver using flat mirrors and can achieve up to Cg ~ 3x. They are very easy to fabricate and capable of giving more output from PV as well as solar-thermal systems. The authors of [29] published that static V-trough collectors provide over an annual average Cg >1.2 in highly diffuse solar fraction conditions, and were found suitable for PV. For convenience, 2 < Cg < 10 is referred to as the low concentration within this thesis.

ii. Concentration Method (Refractive, Reflective, Luminescent)
The key approaches to achieve concentration are reflective, refractive and luminescent, plus total internal reflection which is included within refractive and luminescent types [23]. The fundamental difference between reflective and refractive modes is that the former is based on parabolic reflectors/mirrors (e.g. Cassegrain) while latter work with lenses (e.g. Fresnel lens). Fig. 10 schematically differentiates between (a) reflective versus refractive optics while (b) Fresnel lens versus conventional curved lens.
Cassegrain mirror assemblies are made of a parabolic primary optic and a hyperbolical secondary optic [30]. The mirrors can achieve ultra-high concentrations ranging from 500-1000x and they are beneficial, as they don't have chromatic aberration. Fresnel lenses comprise of discrete concentric prisms patterned on a superstrate, either monolithically or with a separate material layer with the advantage of a high optical efficiency that can reach values up to 86 % [31]. Compared to ellipsoidal concentrators, the large size commercial Fresnel lenses (made of poly methyl methacrylate or silicone-on-glass) cost about 100 US$ each [32]. As compared to conventional lenses, Fresnel lenses use lesser material and have thinner dimensions for providing similar concentration levels as can be seen from Fig. 10 (b). Parabolic troughs are mostly utilised for concentrated solar power using thermal generators and require single axis solar tracking, which is based on the rotation of the concentrator or receiver.
In mirror based systems, the solar cells are normally illuminated from below while in refractive set-ups, the cells are under the light source, which means that the cooling system may not cause any shading issues in case of lenses whereas for mirrors based setups, it's an important design consideration [33]. The benefits of using a Fresnel lens refractor or a parabolic mirror are comparable as a function of the luminosity of the concentrator, however, the achievable gain is greater for a parabolic mirror than a flat lens but is lower than a curved lens [34]. As Fresnel lenses offer more flexibility in optical design, allow for uniform flux on absorber, and are less prone to manufacturing errors, they are more suitable for PV systems [35]. In fact, due to the reasons mentioned above, refractive mode is generally deemed more appropriate for PV while reflective optics find their applications in solar thermal systems.
(a) (b) Figure 10: Schematics of a: (a) parabolic trough (reflective) solar collector versus (b) conventional lens and Fresnel lens (refractive). The Fresnel lens thickness (Xf) is much lesser than that for a curved conventional lens (Xc) for the same concentration.
The third category of concentrators falling under this sub-category are the Fluorescent / Luminescent Concentrators (LSC), which are capable of concentrating both the direct and the diffuse radiations without any tracking systems. Weber and Lambe [36] proposed LSC for PV applications in 1976. LSC consists of a transparent sheet doped with luminescent organic dyes to absorb sunlight and focus it on to the solar cell using the principle of total internal reflection. Their efficiencies are limited because of photon emission within the escape cone, re-absorption of emitted photons by the dye and the matrix absorption [37].
Quantum dot concentrators comprise of quantum dots seeded in a plastic material suitable for building-integration and capable of using both direct and diffuse solar radiation [38].
A typical example is an organic solar concentrator.
iii. Image Formation (Imaging and non-imaging) Imaging concentrators are classified as the more conventional optics such as parabolic reflectors/mirrors or Fresnel lenses in which the formation of an image is a necessary or fundamental requirement. Non-imaging concentrators (anidolic optics) on the other hand, refer to those concentrators that collect the radiation on a larger aperture area and focus on to a smaller area, which is wider compared to a focal point, without forming an image of the light source. Proposed by Hinterberg and Winston in 1965 and grown to popularity with design modifications in the early 1970s, CPC are a classic example of non-imaging concentrators [39]. CPC can collect a wide proportion of the available direct and diffuse radiations. A schematic of a CPC is given in Fig. 11(a) and the working in Fig. 11(b). The term non-imaging has often been referred to as a misnomer as the non-imaging concentrators are also imaging devices. Nevertheless, the image formation is not required or the quality of image at the exit aperture is not of much importance, although image formation is not necessarily excluded [40]. Non-imaging concentrators obey the edge-ray principle, which states that non-imaging devices can be designed by mapping the edge rays from the source to the target edge, or if the source boundary rays can be directed to the edges of the target, then all the rays in between will also be directed to the target area [41]. As these designs offer more simplicity and superior collection efficiency in terms of concentrating the incident radiation much greater than the imaging systems, ideally non-imaging concentrators are preferred especially for stationary applications. Other advantages include uniformity in radiation on flat absorbers, widest acceptance angles and economic benefits in terms of no investment in tracking costs such as installation, operation and maintenance [39].

v. Geometry of Optics
On the basis of geometry (shape, size), the concentrating elements can be classified as parabolic, hyperbolic, elliptical, dish, trough etc. The optical designs specific to this research are parabolic and elliptical, the details of which are discussed in later chapters.

vi. Miscellaneous bases of classification
In addition, concentrating systems can be classified based on tracking mechanism such as stationary (without applied tracking), 2D (linear or single axis tracking), 3D (point or dual axes tracking) and quasi-static. It can be further categorised on the basis of secondary optics designs, homogeniser and conic reflector etc. [23]. The two primary concentrating elements used within this thesis are based on non-imaging 2D LACPC (Linear Asymmetric Compound Parabolic Concentrator) and 3D SEH (Square Elliptical Hyperboloid) concentrator. Both the concentrators provide low concentration (Cg < 10x) and hence they aren't dependent on solar tracking and require no active cooling. They offer a wide range of acceptance angles and can be used as stationary concentrators. Also, CPC have other advantages over parabolic reflectors in that they require less precise tracking due to flat optical efficiency response, in comparison to Fresnel lenses they offer higher optical efficiency (>90 % with advanced reflective films or coatings), the smoothness in the CPC surface minimises manufacturing defects and the option to filter portions of the solar spectrum helps reduce cooling requirements for solar cells [42]. SEH concentrators, on the other hand, offer high transparency allowing daylight penetration into buildings. The details on the concentrators used within this thesis including their designs and fabrication are covered further in section 3.2.

Pros and Cons of CPV technology
As with any technology, CPV has its own advantages that come with a few limitations, which are described in Table 1. In addition to limited niche market and high risk investment, the requirement for solar tracking comes at the cost of increasing system bulk, capital, operational, electricity and maintenance costs, which seems like a necessary evil for high and medium concentration systems. Their costs may exceed the PV costs, they may be difficult to install especially in residential applications and may reduce the CPV lifetime due to mechanical wear and tear [43].

BICPV systems: an overview
CPV, when integrated with the buildings are termed as building-integrated CPV or BICPV. The many benefits of using BICPV, over conventional BIPV are higher electrical power output, simplicity in recycling the constituent materials due to less cell usage, lesser space requirement, improved use of the building space as well as environmental implications such as lesser toxic by-products from solar cell manufacturing [22] which are summarised in Table 2 as a qualitative comparison and a few examples of BICPV are shown in Fig. 13 (a, b, c).

Challenges
Many drawbacks associated with BIPV and CPV become a natural challenge for BICPV systems to overcome. In addition, in comparison to BIPV, BICPV is often difficult to mount or used as a construction element such as the installation on façades, as the mirrors can prevent light entrance inside the building and the outwardly protruding receivers creating strain on the structure as well as are have unpleasant aesthetics [22]. In addition, higher temperature in linear CPV systems produced due to higher light flux density incident upon the cells is also a major issue [22]. The factors affecting a BICPV electrical power output are the material, type and quality of the solar cells used, the concentration level achieved by the concentrator, the incident solar radiation and the operating temperature [49]. For a given BICPV working under constant irradiance, the temperature increase limits the efficiency of the system. The panel manufacturers measure the efficiency under the Standard Test Conditions (STC) of 1 kWm -2 irradiance at 25°C temperature, whereas the electrical efficiency is significantly affected by the outdoor weather conditions [50].
At present most commercial Si-based solar cells used in BICPV can convert only up to 18-20 % of the available solar irradiance to electrical energy and the rest is converted into heat, increasing the junction temperature [51] unless the heat is efficiently dissipated.
Various studies have been undertaken to experimentally investigate the effect of operating cell temperature over output parameters. In mono and polycrystalline solar cells, the increase in temperature led to slight increase in short circuit current (0.06-0.1 % /°C), but higher voltage drops (2-2.3 mV/°C), and fill factor reduction (0.1-0.2 %/°C) causing the electrical power output to drop by 0.4-0.5 %/°C in one of the studies [52].
In order to analyse the effect of the cell temperature on the electrical output relationship, it's important to understand the underlying mathematical and physical principles. The output electrical current for a solar cell is related to its temperature as given in Eqn. 5 [68,69].
Where I-characteristic current of solar cell, Iph-photocurrent, Io-reverse saturation current, Rs-series resistance, Rsh-shunt resistance and k-Boltzmann's constant, n-ideality factor, q-electron charge, Rs-series resistance, Rsh-shunt resistance, T-cell temperature and Vcharacteristic voltage of the solar cell. It is based on the single-diode model, the most widely accepted approach for circuit-based PV which accurately predicts the I-V characteristics of the solar cells, especially at T >300 K [68,69]. Voc is related inversely to the cell temperature as given by Eqn. 6.  [70].
The authors of [68] proposed the electrical characterisation of PV modules using the Approximate Single-Diode Model (ASDM) simulation and reported that the I-V curves for two different modules ( Fig. 15 (a, b)) at different temperatures follow the same trend as the experimentally obtained results with high accuracy. This also shows that for any module with varying efficiency, the I-V curve always shifts leftwards at higher temperatures. From the above experimental results and numerical simulations, it's evident that the efficiency and output of a solar cell are degraded at elevated temperature. Not only this, the temperature rise has other disadvantages such as being detrimental to the longevity of the silicon solar cells and is a challenge for the effective functioning of the solar cells within safe operating temperatures [69,71]. Amongst the other challenges faced by BICPV are a lack of available models for predicting the output energy yield and calculating the maximum power [72].

WICPV: a type of BICPV
A Windows-Integrated Concentrated Photovoltaic (WICPV) is a special case of BICPV and is based on the SEH concentrators ( Fig. 13(c)) which focusses sunlight onto a smaller area of 1 cm 2 square silicon solar cells, thereby saving more than 60 % of the solar cells against that used for a conventional PV [47]. Being fixed or stationary in nature, these concentrators can be used without the need for an all-day tracking system, thereby saving operation costs. This distinctive SEH optical concentrator design aids the high efficiency of the system and delivers hassle-free power from the window [74]. This WICPV can be integrated into glazing building façades and contribute to (i) daylight penetration through windows due to their transparent nature (ii) low-cost electricity production at the site of use reducing the energy dependency of buildings and (iii) extremely low solar heat gain [75].

Current scenario and future: BIPV, CPV and BICPV
The energy payback time for a PV panel, which is the time required to balance the input energy required for manufacturing the panel or to recover the initial investment, is a function of the geographical location and is normally calculated in terms of time (years).
For example, for Northern European countries, it's calculated as around 2.5 years, for Southern European countries it's less than 1.5 years (1 year for CPV) and for Italy, it's 1 year for multi-crystalline Si modules, thereby implying that PV systems are capable of producing twenty times the energy required to manufacture them, assuming a 20 years lifespan [2]. The culmination of BIPV and CPV gives rise to BICPV, which is the foundation of this thesis, hence their current and future scenarios are discussed here. BIPV market worth increased from US $1.  While the predicted annual growth is highest for Europe, followed by Asia-Pacific region, the USA and rest of the world, the % CAGR is estimated to be the highest for Asia-Pacific to achieve solar concentration and more recently, for PV concentration have taken place ( Fig. 17). Figure 17: The CPV designs: past and predicted future trends on a timeline. The circles in upper half denote the requirements of high accuracy manufacturing and quality materials while in the bottom half, the most reliable versions are shown [23].
Over a period of time, more attention has been paid to improve optical surface structure, employing a secondary or homogenising optical element and towards reducing the path length of light rays with an aim to improve optical efficiency, tolerance and irradiance uniformity [23]. CPV technology seem a commercially viable solution to increase PV efficiency and reduce the module costs as the typical cost of silicon solar cells are about 160 US $/m 2 in comparison to the lenses available at 30 US $/m 2 and the mirrors at 17 US $/m 2 [43]. It is envisaged that the robust, easily manufactured, and based on an Earthabundant material, silicon, the crystalline silicon based PV will become more costcompetitive without subsidies in the US at the module costs of US $ 0.50-0.75 per peak watt, enabling electrical production at a levelised cost of electricity of US $ 0.06/ kW·h, which is comparable with the present baseload fossil-fuel electrical utility plants [79]. CPV has already entered market as a utility-scale option for solar electricity generation since several of the global cumulative 360 MW instalments are beyond 30 MW per site [44].
As per the data available for until 2016, the few remarkable developments in the CPV market and industry include [80]: The comparable economic advantage of CPV systems at present in comparison to conventional PV is still challenging [81]. It demands the optical systems to be more reliable, an increasing number of higher concentration based systems, lesser reliance on costly electromechanical tracking mechanisms and a reduced requirement of cooling systems. This, in turn, needs greater innovation to be incorporated into new designs, search for novel materials and discovery of newer applications.
In the long run, BICPV viability will depend on the relative ease of replacement of structural elements and provide versatile usage with aesthetics plus cost competitiveness over conventional flat panel PV which already has started declining in prices [22]. The decrease in PV prices in the last four decades has positively influenced their market expansion. In addition, the government policies and incentives such as feed-in tariffs in Europe and investment tax credits in the US have contributed to their penetration [82].
As a result, BIPV can overcome the shortcomings generated by the BAPV systems. While selecting the appropriate BICPV system for a building, it's worth considering the given advantages and disadvantages of the available panel types, ease of integration, type of building architecture, aesthetics, costs, building and government regulations, grid connectivity if feeding into the grid is feasible and the available incentives.

Research Questions and Aims
The objective of the current work is to contribute to the growing research in the field of BICPV and aiming to address the thermal management issues. This PhD thesis primarily investigates the effectiveness of passive thermal management of BICPV, specifically using latent heat in PCM and/or nanomaterial enhanced PCM (n-PCM) based on experimental methods. A small section of this thesis is dedicated to study of micro-fins as secondary temperature regulation in tandem with PCM and/or n-PCM and the relative effect of PCM are compared with a sensible cooling media, water.
The following key research questions are attempted to be addressed:


To characterise PCM and n-PCM using X-ray diffraction (XRD), scanning electron microscopy (SEM) and differential scanning calorimetry (DSC).


To study the suitability of these novel materials with respect to their heat transfer characteristics, thermal conductivity, melting point, specific and latent heat.


To investigate the thermal enhancement effect of using multiple passive cooling techniques in tandem such as micro-fins with PCM and micro-fins with n-PCM based on BICPV temperature emulation, for assessing BICPV suitability.
 A brief experimental work to compare the latent heat medium (PCM) with the sensible medium (water) used in passive cooling, for BICPV thermal control.

Scope of research: Features and Limitations
Given the numerous technologies accessible and options obtainable within this research area, it becomes imperative to set the boundaries of the project, define the in-scope elements and clearly distinguish them at the very outset. Within this thesis, the challenges related to thermal regulation of BICPV systems are attempted to address using a narrow spectrum of the potential solutions. A simple grouping of these technologies using a flow diagram will aid an easy comparison of the relevant research areas and distinguish the selected pathways to addressing them (Fig. 18). The salient features are listed below:   BICPV: Only low concentration BICPV were considered for the study.
 Media: passive cooling; most of the experimental focus was on latent storage media with one reference to using a sensible medium, water for the experiment.
 PCM type: the majority of experiments were conducted using organic PCM, due to the reasons discussed in subsequent chapters. However, other PCM types such as granulated forms were analysed for the thermophysical properties using DSC. Figure 18: Overview of BICPV system challenges and available solutions. In blue: key areas of focus while in green are the secondary areas of investigation.
 Mode: passive cooling using PCM formed the foundation of this thesis but some work on combined usage of micro-fins with PCM has also been reported.
 Sourcing: all PCM and nanomaterials used were sourced from the manufacturers.
However, n-PCM was manually synthesised in the laboratory.  Nevertheless, this area can be an opportunity for the future work in this field.


The thesis is mostly based on experimental results and the obtained data and mathematical analysis with some application of numerical simulation. properties, applications, challenges and means to address low thermal conductivity using nanomaterials are discussed. In addition, passive temperature regulation using micro-fins is considered as a potential candidate. This chapter forms the basis of identifying the knowledge gaps within BICPV cooling and the ways to address them.

Organisation of thesis
Chapter 3 provides the details on materials, methods and design and fabrication for the modules used for experiments. The key instruments, devices and auxiliary equipment used during the preparation and running of the experiments along with their technical specifications are described here. Steps by step engineering design, construction and fabrication processes for the BICPV and the PCM part of BICPV-PCM system set-up.
Following on from here, Chapter 4 describes the methodology followed for electrical and thermal characterisation of these various BICPV modules with natural convection and with PCM applied. The procedure for instrument calibration and varying the test parameters is also detailed here. The four key systems studied are (i) an LACPC based BICPV system for studying the effect of PCM and irradiance, (ii) an SEH based WICPV system with Copper tubes for studying the effect of thermal conductivity enhanced PCM containments and sensible cooling media such as water, (iii) a micro-finned thermal system emulating BICPV temperature for investigating the effectiveness of micro-fins and a synthesised n-PCM and (iv) an SEH based BICPV system for examining the effect of n-PCM on BICPV performance enhancement.
The following chapter, Chapter 5, presents the results from the experimentally obtained thermal and electrical response as measured using the various set-ups as detailed in the previous chapter. It also contains results in terms of analysis of PCM thermo-physical properties (using DSC), morphology (using SEM) and some preliminary results for the material composition and purity analysis (using XRD). The numerical schemes, simulation validations and the qualitative and quantitative comparisons between the different experimental set-ups have also been explained in detail. The results in appendices give details on thermo-physical properties from DSC, and morphology from SEM instruments which will be useful for further work in this field.
The final chapter, Chapter 6, gives a general conclusion to the thesis, the lessons learnt during the entire lifecycle of the experimental work carried out throughout the period of the research. The second part of the chapter outlines the recommendations for the future work that can be carried out to enhance the research. This covers many aspects of research ranging from additional means for thermal energy harvesting from stored PCM energy, advanced level synthesis of n-PCM and characterisation, outdoor characterisation of the BICPV-PCM module to account for the ambient air cooling and the practical effects of non-uniformity of the irradiance on the module.

Introduction
Improving electrical efficiency for the low concentration BICPV systems require efficient heat removal, without a lot on investment and without increasing system bulk. The heat transfer rate from a surface can be enhanced either by increasing the heat transfer coefficient (i.e. by using a pump or fan) or by increasing the surface area (by attaching fins). Another way of looking at this is the option of either employing active cooling (such as water pumped cooled systems, which require mechanical or electrical power for operation), or enhancing passive heat transfer, for example, by changing the geometry (i.e. by introducing fins), or by utilising phase change materials (PCM). PCM are thermal energy storage material used for heat absorption, storage and recovery and are often employed in renewable energy systems due to their intermittent and unpredictable nature.
PCM contribute to the applied system for rationalising and uniformly spreading the use of energy over a period of time. Heat removal can be classified as sensible or latent depending upon whether the storage media removes heat in a sensible or a latent form.
Sensible heat and latent heat describe the heat exchanges taking place under the specific conditions described by their effect on a thermodynamic system. Often, the best option to store energy in these systems is to take advantage of the enthalpy of phase change. The resulting system is typically a latent heat based. Various kinds of sensible and latent heat storage material and system options have been employed for BICPV cooling systems in literature. The classification of PV thermal management modes, mechanisms and mediums is presented in the next section.

Thermal management: modes, mechanisms and mediums
The

Modes: Active and Passive
To address the challenges related to BICPV temperature rise, thermal regulation can be achieved via either active or passive modes of cooling, which are also the established methods for cooling electronic devices [83] and PV panels [84]. Passive cooling doesn't require any mechanical or electrical power input for heat extraction from the system, exploiting natural laws such as free air convection and buoyancy, whereas active cooling depends on externally supplied energy to cool the solar cells [85]. The two modes are distinguished in Table 3. Active cooling has been established as an effective means for PV applications by numerous authors through their experimental and simulation works.
To list a few, [86] achieved a temperature reduction by up to 22 °C , by spraying water over the PV frontal surface using water pumping systems, mean PV array power increase in the range of 21 % [87] and by 9-22 % from the reference values [88]. The authors of [84] experimentally proved that actively cooled PV cells, using an air blower, showed an increase in electrical efficiency from 8-9 % to 12-14 %. Using a parallel array of ducts with inlet/outlet manifold for uniform airflow distribution at the rear side of the panel, the temperature dropped from 68 °C to 38 °C with active cooling. However, the additional costs of operating the blower or pump for active cooling are not reported and could be higher than the electrical efficiency improvement produced.
Previously, passive cooling was not considered feasible for densely packed cells or for linear concentrators with Cg > 20. However, in recent years, passive cooling technologies have become more acceptable for CPV applications as they possess greater reliability and are safer than forced flow, which has a higher damage probability caused by the active cooling failures [100]. A comparison between active (forced air) and passive cooling (PCM) modes by the authors of [101] revealed that passively cooled systems showed better performance at high discharge rates, high operating temperatures and ambient temperatures of over 40 °C without expending significant fan power, in cooling a Li-ion battery pack used for plug-in hybrid electric vehicle propulsion. With PCM, the cell temperature remained below the upper safety limit of 55 °C in high constant-rate discharge with an ambient temperatures up to 52 °C. This was not feasible using active cooling due to the high airflow rate demand close to the turbulent range that is beyond practical range for vehicular applications. However, the authors suggest that in this case, the expense of electric power for running the fan was relatively small.
Other reported passive thermal management methods include silicon cell cooling using immersion in an appropriate isotropic liquid dielectric with high refraction factor, dielectric constant and specific resistance such as glycerine, which has shown to increase the solar cell efficiency by 40-60% [102]. Using thermoelectric technology, PV heat sinking by thermoelectric module tile has shown an increase in electrical efficiency by 1-18 % and panel temperature reduction by 6-26 % for a range of irradiance from 800 to 1000 Wm -2 within 25-45 °C based on Matlab and PV modelling [103]. The applicability of these methods for BICPV is yet to be confirmed.
Within the thesis, keeping in view that BICPV is based on low concentration concentrators, Cg <10, passive cooling is selected as the primary mechanism as it seems to offer a convenient, cost-effective and simple way of temperature reduction. Therefore, passive cooling with PCM will be discussed more elaborately hereafter.

Mechanisms and Media: Sensible, Latent and Thermochemical
Heat removal mechanism can be classified as sensible or latent depending upon whether the storage media removes heat in sensible or latent form and they are described below.

Sensible medium
Sensible medium remove and store the heat energy by means of raising the media temperature such as in dwellings and building structures, where the thermal mass of the materials act as a heat store. The commonly used sensible heat storage media are water (in hot water tanks), ground/soil/earth, bricks, concrete and rock (in rock beds) systems.
Traditionally, energy was stored in the form of sensible heat which required large volume of storage material, while now more emphasis is being laid on latent form in PCM [104].
The sensible heat in a system can be expressed in terms of Eqn. 7 [105].

= (7)
Where Qs is the sensible heat storage (J), m is the mass of the material, Cp is the specific heat capacity of the material at constant pressure (J/kg/K) and ΔT is the change in temperature (K).
The drawbacks of sensible media are: incapability to remove heat at constant temperatures and it is a less efficient method because it takes more energy to change the state of a material compared to raising its temperature by a reasonable degree. They also require significantly larger quantities of media for removing the same amount of energy as in a latent [106].

Latent medium
Latent heat removal mechanism, with media such as PCM is preferred where higher storage densities with a smaller volume of the material (with high latent heat) are required [107]. Operational advantages using PCM over a sensible heat removal system include smaller temperature fluctuations, smaller size, lesser weight per unit of storage capacity and high energy density (typically 5 to 14 times of that based on sensible media for a given working temperature range) meaning lesser area requirement, more reliability and flexibility [104]. In real-life scenarios, the latent heat changes are accompanied by sensible heating, before and after the phase change, hence Eqn. 8 gives the mathematical expression for the phase change. However, the sensible heat effect is often negligible compared to the latent heat during phase transformations, therefore, to calculate purely the latent heat Eqn. 9 [105] can be used.
Where Ql is latent heat (J), Ql, overall is overall heat storage during phase change, ΔTs-l is the solid-liquid phase change temperature, (K), ΔTl-s is the liquid-solid phase change temperature, (K) and H is the latent heat capacity of PCM (J/kg). The key differentiating points between the two mechanisms of heat removal are highlighted in Table 4. Table 4. Distinction between sensible and latent heat transfer (data source [108][109][110][111]). Though the high energy density of latent media reduces the heat losses to the surroundings due to the reduced storage sizes, the authors of [112] concluded that it was uncertain whether the energy density benefits offered by the latent materials were useful for a typical solar cooling application. However, higher annual collector performance was observed with the use of sensible media [112]. In another study, [113] reasoned that there were no significant differences between the performance of the two and though PCM did not prove beneficial in terms of either the efficiency or the cost in exergy to be supplied to the store or thermal power characteristics, they could still be preferred due to their compactness. It was established by [114] through experimental analysis that latent systems were a viable option for solar heat energy storage and that it could be utilised as a substitute for domestic solar sensible applications.

Thermochemical medium
Chemical storage mechanisms rely on the phenomenon of chemical reactions (reversible reaction between two or more substances) or thermochemical sorption using endothermic and exothermic cycles [108]. The underlying principles of sorption are further based on adsorption, where gas bonds with the solid surface without creating new compounds, and absorption, where a new material is formed due to interaction. In absorption, the energy storage density depends on the concentration of the solution whereas in adsorption, the solid/gas medium energy storage density is determined by density of the salt, composite porosity and additives. Generally, absorption couples such as LiCl/H2O, NaOH/H2O, CaCl2/H2O, and LiBr/H2O are employed for absorption while Hydrates and Ammoniates are used for adsorption [108]. Thermochemical energy storage offer high energy storage, low heat losses, capability to conserve energy at ambient temperature as long as desired and they seldom suffer from heat loss problems [108]. For the purpose of this thesis, only latent media are described in detail.

Micro-fins
Micro-fins, intended as extended surfaces for improving heat transfer, have at least one micro-scaled dimension. They have been effectively integrated with optoelectronic systems [115,116], with condensers or evaporators for cooling systems and air conditioning devices [117] and with tubes in brine coolers [118] etc. to enhance heat transfer rate. Micro-fins have proven to offer better thermal performance as well as higher mass specific power. Indeed, they have been found to provide up to 50 % higher power to mass ratio in contrast with the conventional heat sinks [92]. Fins also provide larger heat dissipation capability which means the size of heat exchangers could be more compact entailing less refrigerant load in a cooling system. Use of micro-fins has been preferred to relying on forced convection as they are noiseless, require no power for operation and offer less expensive alternative [119]. Even though they may not always increase the heat transfer substantially, micro-fins prove beneficial in terms of material usage and can be considered useful for the applications dependent on weight minimisation of heat sinks [120], such as with BICPV systems.
The fundamental mode of heat transfer in a micro-fin is via convection but it has already been demonstrated that the contribution of radiation should not be neglected [121,122]. correlations. In their work, the authors of [125] applied microscale heat sinks to a processor chip of micrometer scale. The authors of [126] reported the advantages of micro-structured roughness on heat transfer performance of heat sinks, cooled by forced air and showed that 20% enhancement is observed for finned heat sinks compared to milled ones. The application of micro-fins for thermal management of electronic systems has been studied in [119, 127,128] as well as micro-channel heat sinks for reducing thermal resistance for power electronics cooling [129][130][131]. Application of fins [51,132] and micro-fins [92,133] for high CPV system passive cooling have been already proposed. However, for this work, the notion has been extended to potentially use microfins with PCM/n-PCM for the thermal regulation of low concentration BICPV systems.

Combined usage of passive-passive cooling
In their review on solar cell performance enhancement with the use of various cooling methods, the authors of [134] summarised that PCM layer was found to effectively increase the electrical efficiency of the PV cells and fins on back side of the module have similar effect, with the heat transfer rate determined by the exposed area and wind velocity. The authors of [75], numerically characterized the PCM melting in a plate fin type heat sink focussing on parameters such as aspect ratio, Nusselt number (Nu), To analyse the impact of fin geometry (fin-length, fin-ratio and the angle between adjacent fins) and outer tube thermal conductivity on PCM melting process in a sleeve tube with internal fins, [138] carried out a detailed numerical study, with 2D assumptions.
The produced numerical model fairly agreed with other published experimental and numerical results. The results indicated that the PCM melting time could be decreased with small fin-ratio to a certain extent and the angle between adjacent fins had only slight effect on the melting, while the outer tube conductivity proved to greatly impact the melting process whether or not natural convection was taken into consideration. Further, reducing fin-ratio to increase melting speed did not show any notable effect while the presence of natural convection had highly significant influence. Also, 60-90° angle between neighbouring fins in natural convection condition exhibited the highest effectiveness. The authors of [139] illustrated a new cooling technique for low concentrated photovoltaic-thermal system, using a micro channel heat sink with different volume fractions of Al2O3-water and SiC-water nano-fluids. At different concentration ratios, the effect of cooling mass flow rate and nanoparticles volume fractions, system performance was investigated. Results from the numerical simulation of the thermal model showed significant decrease of solar cell temperature at higher concentration ratio by using nano-fluids compared to using water. SiC based nano-fluid performed better in terms of temperature reduction compared to Al2O3 based, and nanoparticle volume fraction was found to be directly proportional to the achieved temperature reduction and hence electrical power increment. Consequently nano-fluids contributed towards higher electrical efficiency at higher concentration ratio, compared to water. For instance, for high solar concentration ratios where the cell temperature reduces to 38 °C, and electrical efficiency improves up to 19 %. It was surmised that nano-fluids increase thermal efficiency for low (less than 17.8) concentration ratios, while reduce efficiency for higher concentration.
From the literature survey, it has become quite evident that the combined use of passivepassive thermal management techniques has proven effective for solar cells and with BIPV systems. However, no mention has been found on their applications for BICPV systems in literature so far. The higher temperature and concentration of temperature around smaller areas makes the thermal management in BICPV designs different from BIPV .Therefore, one of the knowledge gaps that is addressed during this research work is proposing a combined (micro-fins with PCM and n-PCM) passive cooling technique for the temperature regulation of low concentration BICPV systems.

Phase Change Materials
PCM, as latent heat storage systems absorb and release thermal energy while undergoing reversible phase transformations over a narrow temperature range or at nearly constant temperature [140] thus, their heats of fusion can buffer temperature variations [141]. A schematic ( Fig. 20) shows the working of PCM in a solid-liquid-solid type of phase transformation using the exothermic/endothermic cycle. Within the last few decades, a number for authors have reviewed the importance, uses, and challenges, encountered with PCM use [141][142][143][144][145][146][147][148][149]. The authors of [107], with their broad review of energy storage methods, investigated and analysed latent heat storage materials (thermal, physical, kinetic and chemical properties and economics) such as PCM. They categorised PCM as group I (most promising), group II (promising) and group III (less promising; with insufficient data) based on properties such as the Tm and the H and the number of carbon atoms. Measurement techniques for Tm and H, based on differential thermal analysis (DTA) and differential scanning calorimetry (DSC) using alumina as the reference material were also discussed in brief. presented. Numerical simulation of latent heat storage systems, enthalpy formulation and numerical solution for the moving boundary or Stefan problem was suggested and solved using algebraic equations using the control volume technique developed by Voller [150] and Patankar [151]. An interesting concept of off-peak electricity storage was proposed in [152] wherein PCM were melted to stock surplus electricity as latent heat energy providing hotness/coldness when required, which can reduce peak load requirements and uniformalise electricity demand, thereby achieving cost reduction.
Selection of PCM could be broadly based on the following criterion, (a) phase change temperature range falling within the desired range [153]; b) high latent heat, specific heat and thermal conductivity; c) low volume expansion and low/no subcooling during freezing [153], d) non-poisonousness, non-corrosiveness, non-flammable, non-explosive and chemically stable [153]; and e) low-cost. It also includes parameters such as full or partial storage availability, freezing and melting heat transfer characteristics, cyclic duty and reliability [154]. PCM are also selected by their charging/discharging rates, heat exchanging surface, thermal conductivity (k) of heat exchange container material and effective k of the PCM [149].

PCM types
 Based on their initial and final states during phase change or physical/chemical transformation with absorption or release of heat PCM can be classified as solid-solid, solid-liquid, liquid-gas and solid-gas types (Fig. 21). Typically, solid-liquid or crystalline solid-liquid solution types are mostly used due to the ease of working and functionality.
The liquid-gas and solid-gas types of PCM are not preferred because of their dependence on compression and low volumetric heat capacity [155,156].  Table 5). Organic PCM comprise of either fatty acids or paraffin waxes (n-alkanes with the general chemical formula CnH2n+2; n = number of molecules). They are more commonly used due to their high heat of fusion, availability in large temperature ranges, low super-cooling, inertness, stability, non-toxicity and relatively low cost [157][158][159]. However, on the flipside, they are flammable, exhibit low thermal conductivity and volumetric latent heat capacity and practically cause leakage issues in molten state due to thermal expansion. Commercially available organic PCM have phase change temperature ranging from -9 ºC to 150 ºC. Inorganic PCM, on the other hand, are based on salt hydrates (general chemical formula AB.nH2O; n = number of water molecules) and they have high volumetric latent heat capacity, higher k than organic PCM and higher heat of fusion [160]. Their limitations, however, are super cooling, easy and quicker decomposition and highly corrosive nature. Commercially available salt hydrate based PCM can be applied for applications with -21 ºC to 120 ºC phase change temperature. In their extensive review, [161] have shown that organic solidliquid PCM have much more advantages and capabilities than inorganic PCM but do possess low k and density as well as being flammable. Eutectic mixtures are made of organic or inorganic compounds [162] and are capable of changing phase between (-114 ºC to 0 ºC). Table 5 highlights the advantages and disadvantages of PCM under these classification categories.  Based on encapsulation size, PCM areclassified as macro (> 1 mm), micro (0-1000 μm) and nano (0-1000 nm) encapsulated [156]. Classification can also be based on the encapsulation container geometry (spherical, cylindrical, rectangular etc.).
Last five decades have seen the utilisation of paraffin waxes, hydrated salts, fatty acids and eutectics of organic/inorganic compound as PCM in a widespread range of applications and Farid et al [153] reason that the availability of PCM in with widely varying melting and solidification temperatures enable their use in extensive applications.  The details on synthesis, characterisation, and their thermal and chemical properties are elaborated within the publications for fatty acids based PCM [164][165][166][167] and paraffin based PCM [168][169][170][171]. Environment friendly and biodegradable bio-based PCM such as soybean, palm and coconut oil [172] etc. they are and owing to their substantially less inflammable nature, are becoming a preferred option to paraffin based PCM which are generally more inflammable.

PCM applications in literature
The wide-spread practical applications of PCM include residential buildings [173,174], under concrete pavements [175], air conditioning systems [176,177], thermo-regulated textiles [178,179], PV panel cooling [134,180], solar dryers [181,182], solar chimneys [183,184], smartphone and electronic device cooling [185][186][187], internal combustion engines [188], electrical super-cooling mitigation, battery thermal management [189] vehicle component thermal buffering [190] etc. Looking into specific PCM applications for thermal management of buildings, intended for higher energy efficiency via integration with building masonry [191], PCM are used as follows: inside buildings walls [192,193], for under floor heating, they may be used in macro-encapsulated form into the ventilated façade in its air cavity for day lighting and room heating [194], for low energy and free cooling of buildings [89,195], as lightweight thermally activated ceiling panels [196], as moving PCM curtains integrated with thermally effective windows [197], to impart improvement in hot water heat stores with stratification [198], as tiles for building use [199], floor supply air conditioning system using granulated PCM [200] etc.
PCM were also used for capacitance in the air conditioning systems for reducing fluctuations in the daily cooling load, which later developed into direct integration with the refrigeration systems to save energy and for better control [201] and thermal energy storage for solar water heating applications [202].

PCM for BIPV cooling
Recently emerging concept of PV-PCM system integration for temperature control offers an opportunity for extending its usage to BIPV and BICPV systems. Employing PCM passively can retain BICPV temperatures within safe operating limit and can collect the rejected heat for possible regeneration. Literature mentions hundreds of theoretical and experimental evidences and investigations to ascertain the effectiveness of BIPV-PCM and BICPV-PCM systems, a gist of which is presented here.
In their extensive review on PV-PCM, [203]   PCM use for thermal regulation of BIPV has been experimentally reported [207], for a PV-PCM system that achieved maximum temperature reduction of 18   In another relevant work, [208] investigated the commercial PCM's potential for the annual outdoor performance enhancement of a BIPV module outdoors using a simplified heat balance model to calculate the extra energy gain. They inferred from the theoretical study that an increase of up to 3 % of the total energy output with the use of PCM can be expected, however, the additional material and peripheral costs made the system seem economically unviable. They argued that for realising the standard payback period of 10-20 years, the PCM heat capacity may need to be increased by about one order of magnitude; an unrealistic task. Nonetheless, the double role-played by the PCM in controlling BIPV temperature as well as the building temperature made it a more feasible option. In order to investigate the temperature regulation effect of PCM on a rackmounted PV panels [209] experimentally evaluated the heat dissipation in PCM containments with four different configurations.

PCM heat utilisation
As with any other technology, the future of BICPV-PCM integration depends on its technical feasibility, which is proven from the experimental and simulation results already published and the financial viability, in terms of costs incurred versus financial benefits realised. One factor that could support the viability factor, would be the possible utilisation of waste heat from PCM or enhanced thermal energy harvesting. As suggested by Enibe et al. [210], PCM can be utilised for passive solar powered air heating system.
Another way of utilising the heat energy stored in PCM could be to combine a BICPV-PCM system with a solid-state thermoelectric (TE) technology based on either Peltier (cooling) or Seebeck (power generation) mode, which has been reported by [211] for low concentrator photovoltaic. The authors used a novel integrated CPV-TE PCB-based receiver based on Peltier module, which reduced the operational CPV cell temperature by 66.4 °C and increased the absolute cell efficiency by 2.38 %. They concluded that the TE technology proved convenient, reproducible with the capability to almost instantaneously control the operational steady-state cell temperature for different irradiance levels. In a study conducted on the direct cost benefit of the PV-PCM systems by the authors of [212], it was found that the investment costs for mass production of such systems were higher than the financial benefits, thereby reducing appeal in terms of cost effectiveness in cold European countries such as Ireland. However, the results were contradictory and positive when analysed for the warm Asian countries because of the benefit exceeding the costs almost two times, which reassured the feasibility and attractiveness of PV-PCM technologies. The experimental results, numerical simulation and reviews discussed in this section underline the effectiveness of PCM in electronic components cooling, for regulating PV and BIPV temperatures and pave a way for further introductory application into BICPV systems.

Challenges with PCM
The drawbacks of using PCM include low thermal conductivity (of the order of 0.2 W/m.K) [213] in organic PCM, high volumetric changes associated with the solid to gas and liquid to gas phase transformations [214], undesirable change in material properties due to cycling, phase segregation and sub-cooling [215] and corrosive effects [216][217][218] of inorganic PCM. The challenges mainly faced by the organic and inorganic PCM are discussed here.

Thermal conductivity
One of the main challenges with organic PCM as heat removal medium is their low k, which causes delay in PCM phase transformation, leading to slow charging and discharging rates and consequentially challenges in temperature control [219]. A possible solution to address this issue is the incorporation of thermal conductivity enhancers (TCE) [220,221] [226][227][228][229][230], single and multi-walled carbon nano-tubes [231][232][233][234][235]. In addition, graphite in the form of powder, expanded, exfoliated; matrix etc. has also been cited [236][237][238]. Composite PCM with fillers also prove to be very thermally conductive [239]. in composite PCM was noticed using 2%, 4%, 7%, and 10 % mass fraction respectively compared to plain paraffin wax [241]. Preparation and properties of TCE based on Mg and Ca can be found in [242], while thermal performance of PCM with porous media as TCE has been mentioned in [243]. TCE use with PCM for cooling of electronic components has been reviewed extensively in [220] while experimentally studied in [244]. The authors of [245] experimentally analysed TCE for solar chimney applications [246]. Review on TCE specifically for paraffin waxes can be found in [247].

Leakage issues: PCM Encapsulation
A critical issue with solid-liquid organic PCM is leakage [257,258] in liquid phase due to volumetric changes in expansion. Encapsulation or shape-stabilisation is the process to integrate PCM into supporting material as macro or microencapsulating PCM could potentially avoid PCM leakage. The shape-stabilised PCM can be categorised as composite PCM or microencapsulated PCM (MEPCM) [259] and in both the cases, the shell stays stable during PCM phase change (Fig. 23). High-density polyethylene (HDPE) material, formed stable via cross-linking, has been frequently employed as a PCM supporting material due to its high structural strength. MEPCM comprises of a PCM core and polymer or inorganic shell to maintain the shape and prevent PCM from leakage during the phase transformation [260]. This allows greater heat exchange ability because of high amount of heat storage by MEPCM particles [261].  An important parameter, core-to-coating ratio affects mechanical and thermal stability while thermal conductivity of the shell material decides the heat transfer rate. Though encapsulation of metallic shells is difficult, they offer high heat endurance and HT rates, for e.g. silver nanoparticles show substantial improvement in thermal and structural stability of microcapsules on addition with the shell surface [261]. A technique proposed for using PCM as thermal system comprised of forming a two-phase fluid by mixing with another fluid, most commonly, water with PCM.
The five main types of these mixtures [263] are: (i) Ice slurries, (ii) micro emulsions (dispersion of PCM, water using an emulsifier), (iii) Microencapsulated PCM slurries (microencapsulation of PCM in polymeric capsule, dispersed in water), (iv) Clathrate hydrate PCM slurries (clathrate hydrates composed of water as a host molecule, forming a weaved structure where the molecules of other substances are accommodated as guest molecule). The accompanied heat exchange due to the chemical reaction of formation and dissociation of clathrate hydrate amounted higher greater compared to melting of ice into water) and (v) Shape-stabilized PCM slurries (organic PCM infiltrated in high density polyethylene, with a melting temperature higher than of the paraffin). This way paraffin is retained inside the structure of high-density polyethylene, avoiding leakage.
The authors of [264] proposed a BIPV integrated with MEPCM and performed the parametric numerical simulations for thermal and electrical characterisation of the BIPV-PCM system. The results showed that after operating for a day in summer, more than half the m-PCM layer remained in liquid phase while increasing the minimum efficiency by 0.13 %. In winters, a portion of m-PCM did not freeze in the middle, increasing minimum efficiency by 0.42 %. During the normal conditions, the respective increase and decrease in the minimum efficiency was 0.09 % and 0.18 %, respectively for aspect ratios of 0.277 and 1. They concluded that employing appropriate m-PCM layer improved both, the thermal and electrical performances of BIPV and that the performance also relied on the melting temperature and aspect ratio of the panel. Studies in [203] show the potential of the emerging PV/T technology with MEPCM.

Corrosion
Another challenge with organic PCM as thermal energy storage is the corrosion of metal alloy to all the hydrated salts tested. It was also concluded that copper had a corrosion zone that did not increase after extended periods of time [265]. An image of a corroded stainless steel container encountered during high temperature inorganic PCM is shown in

Phase segregation and sub-cooling
Ideally, the latent heat of fusion and crystallization are constant over cyclic uses for a perfectly reversible PCM. However, differences arises due to subcooling or incongruent melting [261]. Inorganic PCM tend to lose their high storage density after repeated use due to incongruent melting and hydrated salt formation, making the process irreversible and leading to the continuous decline in their storage efficiency [153].

Thermal Conductivity Enhancers: Nanomaterials
In order to overcome the inherent low thermal conductivity of PCM, nanomaterial based TCE can be added to a base PCM. Nanoscopic properties differ significantly from their corresponding macroscopic properties; the extremely minute particle size along with increased surface area display unique property creating vast potential for applications [267,268]. Nanomaterials are defined as materials with one or more dimensions in the nanoscale (10 -9 m) range, while nanoparticles are discrete entities with all three dimensions in the nanoscale [269].   [273] carried out experiments using exfoliated graphite nano platelets (xGnP), known for its chemical inertness and resistance to thermal degradation, seeking improvements in the thermal properties of bio-based PCM. xGnP comprises of several layers of graphene sheets with higher aspect ratio as compared to carbon nanotubes. Using thermal gravimetric analysis (TGA), differential scanning calorimetry (DSC), Fourier-transform infrared (FTIR) spectroscopy and k analysis, they concluded that bio-based PCM, completely incorporated into the porous xGnP structure, resulted in 375% increment in k rendering thermally enhanced bio-based PCM (k=0.557 W/mK) more useful than its counterpart bio-based PCM (k=0.154 W/mK). In a study undertaken by [274], n-PCM were developed using pure paraffin waxes as the base PCM and alumina and carbon black as nanomaterials TCE. The resulting n-PCM with alumina showed the highest latent heat enhancement of more than 10%, while the one with carbon black showed high k improvement of above 25 %.

CuO nanostructures
An important class of transition metal oxides such as cupric oxide (CuO, a 3d metal oxide), has been found appealing due to interesting properties such as high thermal conductivity, critical temperature, semiconducting nature with a narrow band gap (1.4-1.7 eV) which make it useful for photoconductive and photo-thermal applications [275].
The crystal structure of Copper (II) Oxide is shown in Fig. 27 (a) and a Scanning Electron Microscope (SEM) image is given in Fig. 27 [279].
CuO nanostructures are used in water chiller systems [234], antimicrobial applications against a range of bacterial pathogens [280], lithium ion storage [281] and CuO nanosheets as anode materials for electrochemical energy storage [282]. CuO nanostructures also find applications in supercapacitors, solar cells, gas sensors, bio sensors, catalysis, photodetectors, energetic materials, field emissions, removal of arsenic and organic pollutants from waste water [283].

CuO nanomaterial based n-PCM as TCE
In [284] in the suspension and the effective k of the nanofluid and that a decrease in nanoparticle size initiated increased Brownian motion of the particles, leading to more particle-toparticle interactions. The various underlying phenomenon considered were the lowering of fluid viscosities and the interaction of fluid with the nanoparticle surface, which produced a shell of ordered liquid molecules on the particle surface that transmitted energy (via phonons), to the bulk of the fluid. This energy transmission resulted in greater k of the suspension so formed, which was measured experimentally to compare with a variety of models. However, the models could not predict the thermal conductivities of the nanoparticle suspensions effectively.

CNT
CNT belong to the structural family of fullerene and are carbon allotropes of seamless cylindrical structure with diameter ranging from one to a few nanometers [290]; some pictures are shown in Fig. 28 (a-d) and SEM images in Fig. 28 (e-g). CNT length size in the range of 100 nm to several cm [291] and they may be open or close ended. Their thermal conductivity depends on the nature of helix and they are good conductors of heat in axial direction while behaving as heat insulator in lateral direction [292]. The span of usage is due to their unique thermal, mechanical, chemical and electrical properties such as extremely high electron mobility (100,000 cm 2 /Vs) [293] leading to high current carrying capacity, high electrical conductivity (104 S/cm) [294], high k (3500 W/mK) [295], hardness equivalent to that of diamond, and high thermal stability. However, CNT as bulk composite materials and thin films based on unorganized CNT architectures have lesser values for these properties than their organized counterparts such as yarns and sheets [291]. Exceptionally high k makes it the most likely choice as filler for composite material. Han et al [296] discussed the effect of nanofiller addition to polymers to increase k. In another study on molecular dynamics simulation carried out by Berber et al [297] an unusually high value of k (equivalent to 6600 W/mK), equivalent of an isolated graphene monolayer or diamond was observed. The reason was large phonon mean free  Figure 28: (a) SWCNT, (b) MWCNT [298], (c) a nanotorus and (d) its continuum model [299], (e), (f) and (g) SEM images of carbon nanorods from different areas (image courtesy Pradip Pachfule et al. [300]).
An experiment by Assael [301] investigated the effect of adding C-MWNT to water.
Graphitic sheets were rolled in the form of C-MWNT with 20-100 annular layersAs dispersant, sodium dodecyl sulphate was used with a 0.6 % (v/v) suspension of C-MWNT in water. They concluded that a maximum of 38 % k enhancement was attained and that the k increase is directly proportional to the ratio of length/diameter which means longer the nanotubes for a given diameter, more is the enhancement recorded.
For the purpose of the experimental work undertaken within this thesis, nano-CuO was selected as a TCE in the preliminary investigation for which a brief literature survey is presented here. The rationale is described in the next chapter. Based on whether the resulting n-PCM improves the thermal performance of the PCM, more expensive TCE such as CNT could be utilised, which offer promising solutions.

Conclusions
There are several possible solutions to address the BICPV heating, out of which passive temperature regulation using PCM is primarily studied. Although apparently sensible media have been implemented in significantly more projects due to the maturity of technology and less expensive material, latent media are preferred for high energy density and nearly constant charging/discharging temperatures. PCM applied to BICPV can increase the overall system efficiency by two means. As a heat sink for dissipating heat, reducing the module temperature and minimising the reduction in BICPV efficiency.
And, as a heat storage system for thermal applications (hot water, space heating and misc. agricultural use), thereby increasing overall system efficiency.
This chapter provides the fundamental evidence and influences the work produced in this thesis, as it introduces most of the theoretical and experimental information that form the foundation of the experimental work presented in the next chapters.  Micro-fins have been used in electronic cooling and light emitting diodes etc. but their application for a low concentration BICPV alongside PCM is also one of the novelties of this proposed research work.
After the literature survey on possible TCE, copper metal filler and nano-CuO were targeted for initial investigations due to availability, low cost and promising results.
However, depending on how effective n-PCM proves for BICPV thermal management, CNT, though highly expensive may prove vastly beneficial. Passive-passive techniques seem to produce further enhancement in BICPV temperature control than the individual techniques applied in isolation. However, these are still in their nascent stage with regards to their application in low concentration BICPV systems. Their viability is only beneficial if the net profit exceeds the additional costs for PCM and micro finning. A suitable option to contribute towards increasing viability is thermal energy harvesting by adding a low temperature TE that has shown promising results in similar experimental proofs without PCM usage. The next chapter will focus on the available and selected materials for producing the BICPV and PCM system components; the methods, design and fabrication process for assembling. In addition, the instruments and techniques used during fabrication and for further experimental analysis are described.

Chapter 3. Materials, Methods and Designs
This chapter deals with the selection of materials, design, pre-fabrication, and engineered fabrication of the BICPV-PCM systems for experimental purposes. It also presents a list of the (i) key materials, (ii) instruments and (iii) analytical techniques, employed to carry out the experimental work contained within this thesis. The BICPV systems were attached with the appropriately designed PCM containments using various types of adhesives.
Joining the internal and the external electrical circuitry followed this and attaching the peripheral devices for monitoring desired parameters.

Methodology
The experimental work carried out within this thesis was focussed on BICPV-PCM systems, hence the material required for fabricating components of (BICPV/WICPV) system (heat source) and PCM system (heat sink) are described separately. Primarily, the

Solar Cells
The solar cells used for this research work were crystalline Si-based LGBC (laser grooved cost, selective emitter structure and low contact shading [17]. The schematic of an LGBC solar cell (Fig. 31), with various layers of metals, alloys and doped metals is given for reference purposes, as there is no further research into this area within the thesis. Similar cells were employed in the original designs for LACPC [46,302] and SEH [47,74] based BICPV or WICPV respectively. It is to be noted that the due to the long-term storage of the solar cells used for this research work, their operational efficiency may have degraded.
However, as the study undertaken within this research is relative or comparative in nature, this would not affect the final results. Figure 31: Schematic diagram of an LGBC solar cell structure; adapted from [48].

(i) Solar cells for LACPC
The 116 mm long, 6 mm wide Saturn solar cells with 6 fingers and 2.5mm bus bar width on the two ends were used with LACPC for BICPV (Fig. 32).
These cells have been optimised to perform under conditions of up to 10 x [75]. Figure 32: LACPC solar cell dimensions (in mm); adapted from [48].

(ii) Solar cells for SEH
With SEH concentrators, NAREC solar cells (11.5 mm long, 10 mm wide) were used with an effective area of (1 cm x 1 cm), hence, nearly square in shape (Fig. 33). Both, the 116 mm long and 11.5 mm long cells, had identical material composition, however, different dimensions. The technical data for both the types of cells are given in Table 7.  surfaces, were bonded to the plain and the micro-finned Al plate using an adhesive backing (Fig. 34). The rectangular heaters provided 28 V, 10.55 W/cm 2 and were supplied with 12 inches of Teflon insulated lead wire.

Concentrators
The concentrators were manufactured using a dielectric clear polyurethane material,  Table 8.

Back-plate
The Aluminium back-plate, formed the base for mounting the skeleton solar cells with the following dimensions: properties), resistance to UV-induced yellow-browning and moisture-induced delamination, stability at higher temperatures and with ultraviolet (UV) light and low cost [304], [305], [306]. A recent interesting report by NREL [307] provided useful information about the types of available encapsulants (Poly-Dimethyl Siloxane or PDMS, EVA, Poly-Vinyl Butyral or PVB, and Thermoplastic Poly-Olefins or TPO and Thermoplastic Poly-Urethane or TPU) and details their thermophysical and chemical properties.
For the experiments within this thesis, encapsulant refers to Dow Corning® Sylgard® 184 Silicone Elastomer (data in Table 9), with the chemical name Polydimethylsiloxane elastomer. PDMS was preferred over EVA due to its exceptional intrinsic stability against thermal and UV light induced stress [308]. Other salient features of Sylgard® are high transparency, good dielectric properties, room temperature and rapid cure processing and easy flow-ability. It's supplied as a two-part (colourless silicone base and silicone resin solution curing agent), clear liquid component, to be mixed in a 10:1 ratio. As per [309], Sylgard® cures without exotherm at a constant rate regardless of sectional thickness or degree of confinement, requires no post cure and can be placed in service immediately following the completion of the cure schedule.    excluding labour costs to fabricate in the laboratory thereby saving a massive 4/5 th costs.

Thermocouples
They exhibit mechanical strength, resistant to oils, acids and adverse fluids and are colour coded to IEC-584-3. Table 11 lists the characteristics of the thermocouples.

BICPV system: Designs
The two key concentrators used within this thesis are: LACPC and SEH concentrators, the design of which are described here.

LACPC design
The design of the manufactured LACPC with a 2-D geometry (similar illumination profile along its length), is given in Fig. 36. The original concentrator design was first published by the authors of [46] in 2004, with the acceptance-half angles as 0° and 50°, offering up to 2.0 x concentration. This was further modified in the subsequent work by [15] in order to optimise the concentrator for building façade applications in the regions of northern latitude (>55° N). The modified design offered higher geometrical concentration of up to 2.82 x with 0° and 55° as acceptance-half angles.

SEH design
With an elliptical entry aperture, hyperbolic profile section and square exit aperture (hence the name Square Elliptical Hyperboloid), these are classified as non-imaging stationary 3-D solar concentrators and work on the principle of total internal reflection. Figure 36: LACPC concentrator: geometrical design [46].
The selected SEH concentrator ( Fig. 37 (a, b)), referred to as H3 in the original design [43,47,74] The rationale behind selecting this particular height of the concentrator was that although H3 has the lowest acceptance angle of 40º (-20º, +20º), it has the highest optical efficiency of 68%. SEH concentrator design and fabrication process has been replicated from as given in [47,74].

BICPV Systems: Fabrication Method
This section describes the basic processes involved in fabricating the different BICPV.

Metal cutting for the Back-plate
The metal back-plate, used as the skeleton solar cells base, was cut from a commercial plain Aluminium sheet roll (99 % pure) provided with a protective cover on both sides to avoid oxidation. The metal guillotine machine was used for cutting the sheets, which strongly held the sheet with the table preventing any lifting during shearing operation.

Wrapping the Masking tape
After the sheet was cut, polyimide based Kapton® tape was manually wrapped around the cut-to-size Al sheet for mounting the CPV assembly on top side. This was adopted to ensure appropriate electrical insulation between the solar cells and the conducting Al back-plate. The plate was set-aside after this.

Soldering the Solar cells
The solar cells were soldered in a series connection (with the front side of one cell connected to the back side to the subsequent cell and vice-versa) to ensure that the current through the connected strings of the solar cells remains the same. The total voltage, however, is the sum of the two voltages for a given current. In addition, series connection uses a single cable and handles larger voltages, hence preferred for connecting multiple solar cells in nearly all PV modules. The soldering process is briefly described here.

1.
Using a soldering machine (power unit and soldering pencil) as discussed in previous section, five solar cells were connected in series for the LACPC (Fig. 38 (a)), eight cells for the SEH with Copper (Cu) tubes (six strings of which were soldered in parallel making 48 cells altogether as shown in Fig. 38 (b)) and six cells for the SEH based BICPV (four strings of which were connected in parallel configuration as shown in Fig. 38 (c)

2.
The skeleton of these connected cells were wiped with a very light, even coating of Dow Corning® 92-023 Primer [310] to strengthen the bonding between cells and concentrator and to improve the inhibition resistance of cured silicones.

Casting the concentrators
The various BICPV modules were based on LACPC and SEH concentrating elements, providing low concentrations in two dimensions and three dimensions (2-D, 3-D) respectively. It is to be noted that both these concentrators were designed within the research group. An array of concentrating elements was fabricated via casting process using Smooth-On Crystal Clear 200® in the stainless steel casting moulds ( Fig. 39 (a) for LACPC and Fig. 39 (b) for SEH). The dielectric clear polyurethane material resin available as a 2-part mixture, was mixed to make the concentrators and the details on casting process based on in [74,311] is synopsised here: 1. The stainless steel casting moulds were cleaned, dried and sprayed with a thin even coating of Universal® mold release agent to easily release the solidified dielectric concentrators from the metal moulds.

2.
Crystal Clear®200 casting resin (A: B mix ratio 10: 9) was manually weighed and mixed in glass beaker and stirred thoroughly to ensure proper mixing.

3.
The mixture was degassed in the Vacuum oven at room temperature for 20 minutes for a few times.

4.
The mixture was gently poured into the moulds to avoid trapped air bubbles.

5.
The mould was left to cure for 24 hours at room temperature, after which the casted concentrating elements were removed using gentle tapping.
Using a single casting mould, 10 rows of individual linear concentrating elements and 3x4 arrays of 12 individual SEH elements were produced. Depending on the experiments, they were adjoined or cut to size using a band-saw for the module.

Encapsulating the solar cells
The solar cells were fixed on the back-plate, using silicone based Sylgard®, which also glued the concentrators on top of the cell assembly. The typical process consisted of:

1.
Mixing: the two parts of Sylgard were poured in a clean glass beaker in the prescribed ratio (10:1), manually weighed on a micro-scale and then poured slowly in another beaker, thoroughly mixing using a clean glass rod.

2.
De-Airing: the mixture in a vacuum oven to remove the trapped air bubbles for approximately 20 minutes.

3.
Pouring: the de-aired liquid was gently poured over the back-plate wrapped in Kapton and with the soldered solar cells temporarily glued on it.

4.
After leaving the encapsulant to cure for about 5 mins, the concentrators were fixed and pressed lightly with a uniform moderate weight for bonding the surfaces and providing optical coupling.

5.
Sylgard® cures to a flexible elastomer requiring no post cure and can be placed in service immediately following 48 hours at room temperature.

Welding the thermocouples
The thermocouples were manually tip welded using a thermocouple welder at the medium level arc and medium level power (Fig. 40). No purge/argon gas was used. The thermocouple tips were attached at BICPV locations using a small piece of strongly adhesive aluminium tape and the twisted wires were reinforced using glue-gun.

Module assembling process
As a general standardised method for assembling the module, the concentrators were attached on top of the series connected silicon cells skeleton using optically clear Sylgard®, poured over in a uniform layer, left for 24 hours to cure at room temperature.
This ensured an appropriate optical coupling between the concentrators and the soldered cells assembly, as an adhesive as well as a protective coating from mechanical damages.
Another thin layer over cured silicone ensured prevention against delamination at higher temperatures. The assembled LACPC module, with 11.9 cm x 8.4 cm concentrator top dimensions is given in Fig. 41 (a). The WICPV module, made of the dielectric material based SEH concentrator (detailed in [47]), had the profile designed using a continuous union between an elliptic entry aperture and square exit aperture where the solar cells were placed. Each SEH concentrator assembly measured 10.2 cm x 10.1 cm considering shrinkage and manufacturing allowance, which was assembled on Al metal strips (1.2 cm wide). Four concentrator assemblies with 12 SEH elements were required to cover the 48 solar cells used within the designed system. (Fig. 41 (b)). For the SEH based BICPV module (Fig. 41 (c) For the micro-finned thermal module, a 30 mm x 29 mm x 2 mm Al plate was used as the heat sink and mounting Omegalux® Kapton flexible heater ( Fig. 42 (a)). Micro-fins were manufactured using micro-milling facilities at the University of Strathclyde (UK) ( Fig.   42 (b)). Al metal was selected to emulate the experiment performed in [311], however, the plate thickness of almost 1.5 times higher than in [311] was selected due to practical challenges on down-scaling. The aspect ratio of 1.1 contributed to the robustness and durability of the component against bending. The shape and dimensions of micro-fins can be found in Fig. 43.

PCM Systems: Materials
This section details the various PCM used as well as the PCM containment designs and their fabrication processes. Costs have been given for reference purpose.

PCM Materials
The PCM were outsourced from Rubitherm ® Gmbh, Berlin, Germany. Mostly, the focus has been on organic paraffin wax based PCM due to the reasons discussed in the previous chapter. Table 12 enlists the supplier's thermophysical properties of the PCM. Due to much higher temperature than ambient within laboratory, working with lower melting PCM (with phase change temperature < 30°C) did not prove practical.

RT28HC
Rubitherm® RT28 HC (melting range: 27-29 °C), was used in SEH system with Cu tubes as lower melting PCM. This is a solid-liquid type organic PCM with higher heat capacity, and supplied as a white, odourless solid with costs at (US $ 11.5/kg).

RT50
Rubitherm® RT50, (melting range: 45-51 °C) was selected to test the effect of melting range of a PCM on its thermal regulation capability. A solid to liquid type PCM, it was used with the SEH system with Cu tubes as a higher melting PCM. RT50 is a white transparent paraffin supplied as small tear-drop pellets form at a cost of (US $ 6.5/kg).

RT42
Rubitherm ® RT42, (melting range: 38-43°C) was used with BICPV and micro-finned systems. It is an organic, paraffin wax material. RT42 was used with also further treated with nanomaterials to investigate the effect on thermal conductivity for another study. A translucent material supplied as blocks, it has a high latent heat of fusion, low supercooling, chemical non-reactiveness, and low vapour pressure, self-nucleating properties, non-toxicity, environmentally friendly attributes and relatively cheaper costs (US $ 6.5/kg). It should be noted that the PCM was recycled for a few studies, which theoretically doesn't change any properties until it exceeds to thousands of cycles.

GR42
Rubitherm ® GR 42, a solid-solid PCM (melting range: 38-43°C), is a granulate which contains PCM within a secondary supporting structure, such as a natural porous mineral particle. Its cost was comparable to other PCM used within this research (US $ 6.8/kg).
Since it doesn't undergo any change in its form, it was used in the micro-finned thermal system study to take advantage of its shape-stabilised property.

RT55
Rubitherm® RT55, (melting range: 51-57 °C) was selected to test the effect of melting range of PCM in the LACPC first with respect to another lower melting PCM system and then against an insulated back. A solid to liquid type PCM, RT55 is a white transparent paraffin supplied as a block at a cost of (US $ 6.5/kg).

Thermal conductivity enhancers
The thermal conductivity of organic PCM was enhanced using Copper mesh in stage 1 and with a copper nanomaterial in the stage 2 for the initial instances of investigations.

Copper mesh
The woven Cu (purity of 99.98 %) wire mesh, 125 mm wide, 0.25 mm thick and weighing 58 gm/m, designed like a tube sock and supplied by The Crazywire Company, Cheshire, UK was used as the TCE in the first instance. Many authors [312] have found that metallic meshes reduce the complete melting time of PCM and that melting initiates from the corner region of meshes. The woven Cu wire mesh designed like a tube sock (Fig. 44 (a)) was cut using scissors and washed before first use to remove contaminants accumulated during production process. This was then filled in the PCM containment in layers so as to fill the entire volume of the containment as shown in Fig. 44 (b). Approximately 170 gm of mesh was used for the set up. They were layered in a parallel arrangement as studies have found [313] layering without interlayer gap resulted in 80% melting reduction time.
The mesh was filled with the poured over liquid PCM, which solidified within the containment, causing the mesh to stay fixed within.

Nano-CuO
Copper (Cupric) Oxide Nanoparticle or Nano CuO with an average particle size of 60 nm was selected as the stage 2 TCE for organic PCM, (properties in   Safe for use, non-toxic [280], whereas Alumina (Al2O3) nanoparticles are proven to exhibit environmental toxicity effect [75] and Ease of availability.
 Non-oxidising and non-reactive with paraffin wax based materials.

PCM Systems: Design
The containment for storing PCM also act as a heat sink for the BICPV system. It is to be noted, that the term PCM containment also incorporates n-PCM containments.

LACPC based BICPV
The inner and outer dimensions of the PCM containment used for the BICPV system with LACPC type concentrator are shown in Fig. 46. This design was based on heat transfer equations taken from literature based study [136,137] and needed further mathematical treatment, which is explained here. This design was based on modified heat transfer equations as expressed in Eqn. 13 and Eqn. 14 given below. As per the available literature [136], the energy balance for a BIPV-PCM system is defined as follows: (a) For TPV, t < Tm, the relation is given by: A ∆t = A(h 1 +h 2 )(T PV,t -T amb )∆t + (T PV,t+∆t -T PV,t )ρC P ∆xA

(b)
The energy balance for the phase transition is: Where A-front surface area of the BIPV-PCM system (m 2 ), Ir-irradiance on photovoltaic Therefore, the modified equations are proposed as follows: (1-ɳ elec )C g ∆t = (h 1 +h 2 )(T PV,t -T amb )∆t + (T PV,t+∆t -T PV,t )ρC P ∆x (1-ɳ elec )C g ∑ ∆t = (h 1 +h 2 ) (T m -T amb )∆t + ρ ∆x These heat transfer equations for BICPV-PCM systems are based on a 3D control volume based on heat flow perpendicular to irradiance assuming no heat transfer takes place in the walls due to low thermal conductivity (0.1875 Wm -1 K -1 ) of Perspex. supported the containment base. Further, the screwed walls were also sealed using a sealant to ensure leak-proof bonding between the edges of the wall [311]. Once the sides and base was glued, it was set to cure at room temperature for 24 hours.

SEH based WICPV (with Copper tubes)
The PCM containments for individual Al strips were fabricated using six Cu tubes of length 32 cm, with 7 mm internal diameter and 8 mm outer diameter (Fig. 47)

Micro-finned thermal system
For the micro-finned thermal system, micro-fins acted as the secondary passive cooling mode, in addition to PCM cooling. The PCM containment for the micro-finned system was 3-D printed, using ABS plastic with the three-step process for model creation and printing [315] as described below: Once the 3D printed part was taken out, a 1.5 mm hole was drilled using drilling machine was allowing thermocouple entry within the containment. The design of the containment and the fabricated part are shown in Fig. 48 (a, b) respectively.

SEH based BICPV
The PCM containment for SEH based BICPV was manufactured using the 3D printing process. The printed component was loaded into basket for a pre-set time and temperature  Fig. 50 (a) and the design is given in Fig. 50 (b). One 3 mm hole was drilled on one side of the containment using the drilling machine, for accommodating thermocouples. All thermocouples were inserted through this hole, which was sealed using the glue gun for a leakage-proof assembly.

BICPV-PCM system assembly
The following sub-sections describe the assembly of the BICPV part (heat generating element) of the various systems with the PCM part (the heat absorbing elements) for the different module types, and details the applicable design data for these systems.

BICPV (LACPC)-PCM
The aim of designing and fabricating this module was to examine PCM effectiveness and the effect of varying irradiance on the BICPV electrical performance. The back-plate of BICPV was attached as the top covering for the PCM containment using acrylic glue.
Thermocouples were attached at the rear side of the top Al plate, entering through the holes drilled in the containment walls, sufficient to accommodate the multiple wire thicknesses. The BICPV and BICPV-PCM system so assembled are shown in Fig. 51 (a,   b) respectively while the properties of the individual key components are enlisted in Table   14. The design data for the mathematical analysis is given in Table 15.

BICPV (SEH)-n-PCM
The objective of fabricating this module was to examine the n-PCM effectiveness for BICPV performance. The system was assembled using the same process), but using SEH instead of LACPC, 3-D printed containment sealed using Loctite® hot melt glue gun which provided better sealing properties quickly. The system is shown in Fig. 52.

WICPV (SEH)-PCM with Copper tubes
The WICPV-PCM module was designed to study the comparative effectiveness of latent (PCM) versus sensible (water) cooling towards improving electrical efficiency. The cooling media was contained within highly conductive metallic copper tubes, fitted at the rear side of the WICPV under Al back-plate with recesses, to ensure that the low k of PCM was overcome by the higher k of Al and Cu, for higher heat transfer. To ensure semi-transparency effect in the module compared to the original design, narrow strips of opaque metallic back-plates were used to mount the skeleton of the soldered solar cell assembly. The remaining structure was made of acrylic based clear plastic frame, so as to ensure provision of daylight penetration within the building. The top view ( Fig. 53 (a)) and the cross-sectional view in (Fig. 53 (b)) can be used to visualise the overall module design. To measure the temperature under the back-plate, six K-type exposed welded tip thermocouples were placed between the Al strips and the Cu tubes using highly conducting thermal paste (thermal conductivity 3.4 Wm -1 K -1 ). These tubes, primarily used as PCM containment, also acted as thermal conductivity enhancers and were insulated using a 3 mm thick glass wool strips between the recesses in the transparent acrylic frame that formed the foundation of the system. This transparency of the frame permits the use of the set-up as a building window or fenestration. Further, for water circulation within this system, a water circulation circuit was designed, using the same plastic pipes, which is described in the next chapter.

Micro-finned thermal system
For the micro-finned module, rectangular Al plates (either with or without the machined micro-fins) were snug fit on 3-D printed PCM containment which was insulated with a polystyrene layer (k = 0.03 W/m·K) on all sides. The system was sealed using silicone sealant to ensure added protection from PCM leakage upon phase change to liquid state.
To ensure the fin contact with the PCM, it was tested after filling the containment for the first time with liquid PCM and upon solidification it was tested again. The heating elements were glued on the top flat surfaces of the plate, acting as a resistive heat source.
The complete fabricated experimental set-up has been shown in Fig. 54 (a, b).  Table 16.   (rounded off to 2 decimal places) in order to receive an irradiance intensities of 1200 Wm -

Differential Scanning Calorimeter (DSC)
Differential scanning calorimetry is used to study the thermal transitions of materials such as melting temperatures and enthalpies (heats of fusion), crystallization and glass transition temperatures. The temperature difference between a sample and a reference substance of a fixed mass heated at a pre-defined rate is used to identify the chemical or physical changes in a material via exothermic and endothermic processes. A purge gas (usually Nitrogen, but also He and Ar) is circulated within the DSC cell to remove moisture, oxygen or dirt particles that may affect the measurements or damage the cell.  The DSC instrumental set-up consists of the main equipment with the furnace, the nitrogen cylinder for providing the protective gas, a data acquisition system (computer running the Proteus software) attached to the DSC equipment and an auxiliary cooling accessory or intercooler unit IC70, which is capable of handling temperature as low as -70 °C and as high as 600 °C (Fig. 57 (a)). A set of standard samples (one sapphire disc each, Ø 4 mm, thickness 0.25, 0.5, 0.75, 1.0 mm) for calibration measurements (specific heat or sensitivity/enthalpy), with purity certificate was provided. The sample cutter and accessories such as a micro-spatula and a surgical blade were used to prepare PCM samples of masses 4-10 mg. The Al Concavus pans and lids had 30/40-μl capacity with outer bottom diameter of 5 mm (Fig. 57 (b)). Proteus® analysis software programs are available as a 32-bit Unicode version, implying that they support the settings which vary by location (decimal separators, data formats etc.). Other features include curve evaluation (added to the method for automatic evaluation of the measurement curve) and specific heat package (to determine Cp for solids, powders and liquids). The software can also be used for carrying out the heat-flow calibrations with sapphire as the standard material. Fig. 58 shows the GUI display, with the DSC curve obtained for a typical thermal analysis experimental run.

Scanning Electron Microscope (SEM)
Optical microscopy is suitable for low-magnification (1000 X) analysis into colour, size and shape detection of samples but for higher resolutions, SEM is useful. The components and working of SEM used for PCM/n-PCM analysis are described here; the two SEM instruments used for the thesis are described as SEM-I and SEM-II. Basically, primary electrons striking specimen react with its surface producing the following beams [320]: (i) Secondary electrons -lower energy electron beam (5-50 eV) produced from the samples after the SEM primary electron beam hits them.
(ii) Back-scattered electrons -same energy primary electrons reflected due to elastic scattering, providing good image contrast in the imaging field.
The key SEM instrumental elements are shown in Fig. 59 (a).
 Electron source or gun creates the electron beam discharged within a small volume with a slight angular spread and predefined energy.
 Lens system containing numerous electromagnetic or electrostatic lenses. It allows the beam from electron source to enter and upon exit, hit the sample surface.
 Scan unit produces onscreen image by modulating the detected system signal upon the beam movement in a raster pattern over the specimen area.
 Detection unit picks the backscattered electrons, secondary electrons and X-rays signals from the sample and transforms them into digital signals. device was 10 mm, which is the OEM provided safety limit as well.

(ii) SEM-II
Hitachi S3200N SEM-EDS (Fig. 59 (c)) was used for high resolution imaging of n-PCM using gold sputtering. It can achieve a magnification of ~ 20 to 250,000 X, with a practical operational magnification of ~60,000 X. BioRad Sputter Coater (Fig. 59 (d)) from BioRad Microscience Division along with the Polaron E5550 density thickness FTM unit (Polaron equipment Ltd., UK, now Quorum Technologies Ltd.) was used to sputter coat the PCM and n-PCM samples in the specimen chamber using nano-layered gold coating.
The thickness of coating was controlled using the Plasma current (electric kV) by means of turning the HV control knob.

X-Ray Diffractometer (XRD)
X-ray Diffraction, an example of X-ray wave interference, is used to produce the periodic atomic structure of crystals. It can be used for measuring the average spacing between the interatomic rows, determining the crystal or grain orientation, discovering the crystal structure for unknown materials and for defining the size, shape and internal stresses for small crystalline areas. The atomic planes of a crystal cause an incident beam of X-rays to interfere with one another as they leave the crystal. The phenomenon is called X-ray diffraction. It works on Bragg's Law as given by Eqn. 11.

= 2 (11)
Where, θ is the angle of incidence at which the cleavage faces of crystals appear to reflect X-ray beams, d is the interatomic layers distance in a crystal, λ is the wavelength of the incident X-ray beam and n is an integer.
For detecting the purity of nanomaterials and PCM GR42, Siemens D5000 Powder Diffractometer, a bench-top X-ray Diffractometer, also configured as a reflectometer, was used for Control software allows automated data collection and visualisation. The specifications are given in Table 17. This XRD is used to examine mineral samples based on crystallography, with a lower detection limit of ~5 %.  Each channel of the instrument has the ability to be configured separately for any of the 14 available measurement functions and provides built-in signal conditioning.
ExceLINX-1A software was used to acquire data directly into an Excel spreadsheet format. The specifications for the instrument are given in Table 18.

I-V curve tracer
Eko I-V Curve Tracer (Fig. 61 (a)) was used in conjunction with MP-160i software ( Fig.   61 (b)) for the CPV module testing and I-V characterisation in combination with the solar simulator, as determined by changing the electrical load during the I-V sweep (Fig. 61 (c)). MP-160 I-V tracer can be used for testing single or multiple solar cell or module characteristics in combination with a solar simulator or natural sun light. The radiant energy on the module can be measured with the voltage and current measurements. The technical specifications for the curve tracer are given in Table 19.   RDXL12SD 12-Channel Temperature Recorder (Fig. 62) with Excel-Formatted Data Logging SD Card was used for measuring and recording the temperature of the various points within the PCM system. As a hand held mobile device, it offers the flexibility for moving and placing, hence more suitable for the micro-finned thermal system, which didn't require a solar simulator. The technical specifications are highlighted in Table 20.

Stratasys' uPrint SE plus 3D Printer based on patented FDM® (Fused Deposition
Modelling ™) technology was used (Fig. 63) to print the PCM containment using ABSplus plastic. The specifications for the printer are given in Table 21.

Power supply
An external power supply (Aim-TTi EX354RD) was used to supply DC electrical power input to the resistive Kapton heaters. Temperature defining voltage and current were set to emulate the temperature rise in the BICPV system as experimentally determined in the literature. The power supply unit came with mixed-mode regulation (to combine HF switch-mode pre-regulation with linear final regulation) and two outputs (0 to 35V / 0 to 4A). The dual output DC bench power supply with digital V, I meters offered the option to work on constant voltage or constant current operation. The power rating was 280W, with minimum and maximum supply AC voltage of 110V and 240V. The calibration for V and I was performed using an external multimeter

Miscellaneous devices 1.
Multi-meter -Fluke® 117 true RMS digital mustimeter was used to measure the instantaneous input current and voltage, which were then used to calculate the electrical power and heat for the micro-finned thermal system with micro-fins.

2.
Ultrasonicator -Hilsonic ultrasonic cleaner with the attached ultrasonic bath was used for mixing and dispersing the nano-CuO particles within PCM. The principle is based on using the ultrasonic energy for agitating the particles.

Conclusions
The various materials employed for the BICPV

Chapter 4. Experimental Characterisation
The

Introduction
In order to produce a fit-for-purpose BICPV module, it's vital to understand the electronic properties of the solar cells that it is composed of. A broad range of techniques is available for characterisation, mainly relying on different electronic structure and the different physical effects. To typify a solar cell's performance, electrical characterisation is performed. This is useful for research and development as well as during the manufacturing process of PV cells and involves measuring the output current or capacitance as a function of an applied DC voltage. Electrical characterisation is important in determining how to make the cells as efficient as possible with minimal losses and a range of key device parameters can be extracted from the DC and pulsed current-voltage (I-V) measurements [322]. I-V characteristics of a material provide information on the carrier concentration and the minority carrier diffusion length [323].
For comparing the different types of solar cells, they are rated on the basis of standardised measurements, usually conducted under standard test conditions (STC), which involves testing the cells using a vertical light source of 1000 Wm -2 , at 25° C ambient temperature and 1.5AM (air mass) spectral composition of the light [324]. The AM factor denotes the path length of the solar radiation passing through the Earth's atmosphere. An AM=1 signifies the shortest light path with vertical solar radiation, and a higher AM implies a small angle of incidence as the light travels longer through the atmosphere. The output measured under STC determines the solar element's rated module output that the manufacturer is obliged to state [12].  It's also important to assess the thermo-physical properties and the phase transformation temperatures (both range and peak) of the PCM. Any inaccuracy in these data may affect the correct choice of PCM as well as the appropriate amount of material required for a particular application. The melting and congealing temperatures and the specific and latent heats of the materials have been validated using a DSC instrument. The morphology and orientation have been assessed using SEM while the purity and composition; using XRD techniques. The procedure for obtaining the DSC thermo grams and SEM micrographs has been detailed in subsequent sections of this chapter, which range from sample preparation to data acquisition and analysis.

BICPV-PCM system: Electrical Characterisation
The I-V characterisation of the modules were performed indoor, using highly collimated  Fig. 65) of the curve tracer. In order to calibrate the simulator and to regulate the irradiance intensity for varying the available irradiance for the experiments, the following processes were followed:

Solar Simulator Calibration
The solar simulator lamp intensity was calibrated using a 20 mm x 20 mm reference solar cell made of monocrystalline Si, enclosed within 9 mm x 70 mm x 16mm anodized Al with a quartz window for protection and temperature sensor. Using a 100 ohm Pt RTD /Type K thermocouple as the temperature sensor, it was used to calibrate the 1 sun (1000 Wm -2 ) irradiance of the solar simulator. The reference cell operated on < 200 mA, between 10°C-40°C [325].

Solar Simulator Intensity Variation
With the Voss electronic GmbH HelioCon device, the simulator's Xenon lamp intensity was controlled based on HelioCon software. For running the device, communication port

LACPC
To study the effect of irradiance on the electrical parameters of the LACPC based BICPV system without and with the use of PCM within the same system, and ascertain the respective correlation, the radiation intensities of 1200 Wm -2 , 1000 Wm -2 , 750 Wm -2 and 500 Wm -2 were selected. The schematic of the experimental set-up along with the peripheral devices and the electrical parameters for the simulator are shown in Fig. 65.
The data was recorded for over 2 hour-period at an interval of 5 minutes. The effect of intensity was studied using the HelioCon control box for regulating the Wacom solar simulator Xenon intensity for both the BICPV and the BICPV-PCM systems. Similarly, for studying the effect of PCM thermal conductivity, the PCM (RT42) was replaced with the PCM and Cu-mesh combination. In order to do this, the PCM container was filled with the mesh, cut using standard scissors and then formed into layers, until the layers were thick enough to fill the entire volume of the containment. Liquefied PCM was then poured over the mesh to form a highly thermally conductive PCM system and PCM was allowed to cool down.

SEH with copper tubes
To test the WICPV system (with SEH concentrator and copper tubes) indoors with natural air convection, PCM and water-cooled options, the solar simulator was used to provide the irradiance. Intensities of 1000 Wm -2 and 500 Wm -2 normal to the horizontal surface were used for 1.5 hours. This limited time period was selected so as to focus on the heat absorption /melting phase of the PCM heating -cooling cycle. Eko® I-V curve tracer MP-160 was used to measure the output Isc, Voc, Pm, and overall system electrical efficiency ηelec was calculated from the obtained Pm data. Using the experimental set-up ( Fig. 66) each experiment was repeated three times for repeatability and accuracy.

SEH
The SEH based BICPV-n-PCM system was illuminated with the solar simulator at 1000 Wm -2 and 500 Wm -2 for a period of 2.5 hours with applied n-PCM passive cooling and for 2 hours for characterising the module without any cooling (Fig. 67).

Micro-finned thermal system for BICPV applications
For this experiment, two rectangular aluminium plates without (for benchmarking) and with plate micro-fins geometries were considered. The aluminium plates (either with or without the machined micro-fins) were fitted on the 3-D printed PCM containment, which was then insulated with a 5 mm thick polystyrene layer (0.03 W/m·K) on all sides. The Kapton heaters were glued on the top flat surfaces of the plate, acting as a resistive heat source. The electrical power input was applied using Aim-TTi EX354RD dual DC power supply. Temperature defining voltage and current were set to emulate the temperature rise in the BICPV system as experimentally determined in the literature [311]. As the maximum temperature attained within the aluminium back plate under the highest solar irradiance (1200 Wm -2 ) was 60 °C [311], for the tests within the present experiment, 70 °C was attained at the selected DC V, I combination (i.e. at P = 0.8W). This is because in this case, the top surface had been insulated to avoid radiative and convective losses, contrary to the uninsulated reported BICPV system, so a higher temperature range was selected to account for the insulation effect. Fluke® 115 service engineers digital multimeter was used to measure the instantaneous input current and voltage from the power supply, which were then used to calculate the produced electrical power and consequently, temperature rise. The experimental set up is shown in Fig. 68.

BICPV-PCM system: Thermal Characterisation
The thermal characterisation of the fabricated modules, without and with PCM, was based on the recorded temperatures using the K-type thermocouples. In general, for all the four experimental systems, welded thermocouple tips were attached to the location of temperature monitoring. The other ends of the wires were either connected to the type k thermocouple plugs, to insert into hand-held temperature recorder or manually screwed inside the 20-slot channels of the temperature acquisition system. In order to calibrate the thermocouples, the standard procedure of using 0°C (ice point) and 100°C (steam point) as the reference points were used. By inserting the welded tip of the thermocouples in a bath, the output from each of them was noted down separately. Only the thermocouples that showed a stable and constant temperature at these fixed points, which also remained uniform, were selected for their suitability for the measurements. The thermocouples were re-calibrated each time before using them again for the experiments. The ambient temperature and relative humidity were monitored using Hygrometer Testo 608-H1 placed in the vicinity of the modules.

LACPC-PCM
Five K-type thermocouples were attached to the back of BICPV using Aluminium adhesive tape for thermal conduction. Similarly, another 5 K-type thermocouples were attached to the inner side of the bottom Aluminium plate of the containment (Fig. 69).  the experimental run with water, an arbitrary flow rate of 5 mLsec -1 was selected, a rate at which the flow was visibly streamlined; below this rate, the flow in the pipes remained constricted and higher rates produced more turbulence, due to high Reynolds number.

SEH with copper tubes
The water circuit for circulating water is illustrated in Fig. 70. Five clear soft plastic tubes on either side of the copper tubes were used as conduits for inter-channel flow. It is to be noted that for these experiments, the difference in heads (of water supply and the WICPV system) acted as the driving force for the water flow, doing away with the need for an electrical motor or pump thereby saving the active power loss.
Hence, it may be argued that although water was continuously circulated for this experiment, it is still a passive cooling mechanism.

Micro-finned thermal system
The schematic of the experimental set-up has been shown in Fig. 68

SEH-n-PCM
For the SEH BICPV module, the temperature was recorded under the back-plate every 10 seconds for a duration of 3 hours, using a handheld Omega RDXL12SD 12-channel temperature recorder. Five K-type thermocouples were placed at horizontal locations; attached to the rear side of the top plate of the BICPV module using Aluminium adhesive tape and secured using the Loctite glue gun (Fig. 71) and five others were located on the bottom of the containment.

DSC analysis of PCM
The working principle of the DSC is based on the measurement of the difference in the amount of heat used to increase the sample and the reference temperature as a function of either time or temperature. However, the results depend on the sample mass as well as the heating rate applied to the sample. It has been reported that the melting enthalpies measured experimentally by the authors of [326] for various PCM samples were lower than the manufacturer's values, with uncertainties ranging between 10 % -47 %.
Typically the data supplied by manufacturing companies are determined for a large amount of samples, which may be slightly or overwhelmingly different when the PCM is used in relatively smaller quantities such as, for the experiments within this thesis.
Therefore, a DSC analysis becomes a necessity for detecting the magnitude of inaccuracy within the supplied PCM properties during phase change. The working on DSC involves a methodical process including sample preparation, calibration, evaluation and interpretation of the resulting curves, which are described below.

Samples
For the experimental work undertaken within this study, a wide variety of PCM including organic (RT series, based on paraffin wax), inorganic (SP series), granulated (GR series) and powdered PCM (PX series) were studied and analysed using DSC. This is done with a foresight of developing combination PCM from within the available different material types.

Sample preparation
The samples for DSC were prepared by sealing 10 ± 0.015mg sample of PCM in an aluminium (99.5 % purity) sample crucible. Known as concavus pans, (diameter: 5 mm, volume: 30/40 μL) these crucibles have a unique geometry with a concave bottom that allows a clearly defined ring-shaped contact zone. The lids were perforated using the standard perforation tool to avoid any pressure building in the sample during phase change. The pans were covered with pierced lids and cold-welded using a manual mechanical sealing press (Fig. 72) to avoid leakage of PCM in the liquid state that also maintains the furnace hygiene. As the samples for measurement were arbitrarily selected from the bulk supplied PCM, at least three samples were tested for each type of PCM in order to ensure accuracy and repeatability in the results.

DSC calibration
The DSC instrument was calibrated for temperature and heat flow/enthalpy/sensitivity measurements, with known reference standards (high purity reference materials with > 99.99 % purity). High purity metals exhibit sharp melting peaks, which make them popular for calibrating the instrumental response against accurate values. For purposes such as for organic materials and polymers, this may not be appropriate as there is an argument for calibrating like with like, however, indium, tin and lead are still widely used and will remain as convenient temperature and enthalpy standards for DSC in the time to come [75]. One empty reference crucible and the loaded sample crucible were placed in Arena Furnace. The reference crucible was always placed on the left furnace and the sample crucible on the right. The temperature program was set using the Expert mode.
The tests were carried out under nitrogen gas as the purge and protective gas with a flow rate of 40.0 ml/min and 60.0 ml/min respectively. Intracooler IC70 was used as the cooling device, as they offer an alternative to liquid nitrogen cooling.

(i) Temperature calibration
For temperature calibration pure metals were heated through their melting temperature range based on the exactly similar conditions (such as atmosphere and heating rate), which were to be used in subsequent measurement runs with the samples. A comparison was then drawn between the observed and the theoretically known melting points of the standard materials, to determine the resulting difference. The calibration shifting by a constant amount in the sample temperature is based on the number of points (one, two or more) and relies on a smooth curve through the calibration points. The DSC curve shows the melting of pure substances as they melt at an exactly defined temperature corresponding to their melting points. This temperature is taken as the onset or the start of the melting process which is defined as the temperature given by the intercept of the extrapolated slope of the melting curve and the continuation of the base line [327]. Fig.   73 shows the onset temperature calibration chart for the DSC equipment used within this thesis, using five substances (C10H16, Indium, Tin, Bismuth and Zinc). Table 22 represents the nominal and experimental temperature and the temperature correction.

(ii) Sensitivity or Heat Flow calibration
As DSC instruments do not record the temperature directly at the sample, but through the crucibles via a thermocouple, this thermal resistance gives rise to a difference in the accuracy of measured and actual sample temperature, which affects the width of the measured peaks. A narrow peak and steep slope suggest proximity with the actual values [328]. The onset slope of the heat flow-temperature curve is used to calculate the value of thermal resistance [329]. Sensitivity calibration is carried out to determine the difference between the theoretically known and the observed values of heat or enthalpy of fusion for one or more standard materials. The ratio of these values is termed as a Cell constant. Fig. 74 shows the sensitivity calibration chart for the DSC equipment used within this thesis, using six substances (C10H16, Indium, Tin, Bismuth, Zinc and Caesium Chloride) as provided by the manufacturer while Fig. 75 demonstrates the physical usefulness of the calibration. Table 23 represents the heat of fusion, the experimental and calculated sensitivity of these substances.

Temperature program /thermal cycle
To ensure the samples underwent same thermal cycle of charging and discharging, the following generic thermal cycle was used: The samples cooled down to 0 °C from room temperature to standardise the beginning. After maintaining a 5 minute isotherm at 0 °C for achieving a uniform temperature distribution across, the samples were heated dynamically at the same rate to 75 °C followed by another 5 minute isothermal heating ( Fig. 76). In the end, they were subjected to cooling cycle (75 °C to 0 °C) at the same rate. Emergency cut-off was set at 95 °C to avoid any accidental superheating failure. The temperature range on the program was selected to ensure that the samples undergo complete phase change and their behaviour could be recorded for the entire range. Each sample underwent three heating and three cooling cycles using the same temperature program and the measurement data was averaged and tabulated. The first thermal cycle was used to ascertain more uniformity and homogeneity within the samples before characterisation. Therefore, the measured data, phase change behaviour and thermal analysis of the data obtained for this research work were based on the second thermal cycle onwards.

Measurement procedure
The DSC 214 Polyma instrument measurement window was opened in Expert mode and the desired measurement (baseline, sample + correction or sample) was selected. For running the measurements on the samples to determine melting temperature or heat of fusion, the following procedure was adopted (Fig 77): (i) Baseline run: two empty crucibles of same mass, with one reference.

Parameters
Material characterisation is imperative to correctly determine the thermophysical properties such as melting and congealing temperatures, and heat of fusion values for the PCM in use. Although within this thesis, only PCM charging side was employed for the experiments, which is mostly related to the melting characteristics, the PCM thermal analysis was carried out with an aim to determine both the melting and the solidification characteristics. Fig. 78 shows typical DSC thermograms for a PCM with the melting curve  The following points were considered while taking these observations:  For this thesis, the melting peak is the actual melting point of the substance.
 The width, i.e., the difference between the onset, and the end temperature of the solid-liquid transition indicate the carbon number distribution of the waxes [331], which are not studied. However, the temperature ranges for both melting and congealing cycles are studied to compare with the manufacturer data.
 The PCM temperature range was determined by the starting (Onset) and ending of melting/solidification throughout which the phase transition was in progress.
 The software of the temperature programme automatically determined the peak as the maximum (most positive) or minimum (most negative) values of the DSC signal during melting and solidification cycles respectively.

(vii) Latent heat (L):
It is important to determine L of PCM because this determines the amount of heat that can be stored. On a specific heat -temperature curve, it is calculated as the area under the curve. Fig. 79 shows total latent heat capacity of the material between 10°C and 50°C (total cross-hatched area of 197.7 J/g) and the peak area (double cross-hatched area as 150.7 J/g).

(viii) Specific heat capacity (Cp):
The Cp Ratio method was selected for evaluating Cp of the PCM samples. The two correction and sample runs were opened on the Proteus Analysis window using temperature scaling for the measurement (on Xaxis). The appropriate dynamic heating segments on both runs were selected and sapp.nbs.cpe was highlighted. The three DSC curves, Baseline, Standard and Sample were selected and Cp curve was auto-calculated in a new window. As can be seen from Fig. 79, maximum or peak Cp was calculated as the value at the tallest peak (25.295 J/g/K). However, the baseline Cp for the material, which is provided by the manufacturers, can be given by the base Cp as highlighted in Fig. 79 (approx. 2.5 J/g/K).

SEM analysis of PCM-I
To study the morphology of nanomaterial samples and to obtain the micrographs of n-PCM samples for studying the dispersion of nanomaterial within the PCM, FEITM Quanta FEG 650 SEM was used. Samples of nanomaterial nano-CuO (60 nm) and n-PCM (RT42 enriched with 0.5 % w/w CuO) were tested. The nano-CuO samples were dispersed in 100% pure methanol magnetically stirred whereas the n-PCM samples were ultra-sonicated again before scanning them to enhance the dispersion and to avoid particle sedimentation thereby ensuring homogeneity in sample. The imaging process based on [320] is described here.

Sample preparation
The specimen was prepared on the stub that fits on top of the sample holder, using a carbon tape for adhering it. The stub was screwed into the end of the threaded rod ( Fig.   80 (a)) after the specimens were filled on the prescribed carbon-coated stub (Fig. 80 (b)).
The surfaces were blown using a nitrogen gun to remove any loose particles, which were then set-aside and inserted in the holder within the main chamber.

Venting and Loading
To minimize the time of exposure of the chamber to ambient air, it was vented while loading and unloading the sample as the contaminants and humidity can affect the quality of the sample images and increase the pump down times.

Imaging
The NavCam (navigation camera) was swung at 90° for activation and the stage was moved under the NavCam to capture the image. During images being taken, the WD was adjusted but a safe distance of 8 mm was ensured.

Venting and Unloading
The beam was turned off and the sample stage was lowered to the 10 mm mark before venting. The sample holder was removed and the chamber door was closed. The high vacuum button was clicked in order to pump the chamber again to high vacuum.  [332]. Therefore, it is advised that the insulating samples be coated with extremely thin layers of a highly conducting material such as Gold or Platinum to improve their operation in high vacuum mode.

SEM analysis of PCM-II
Within this thesis, Gold-sputtering was performed as described below.

Gold-sputtering
On the BioRad Sputter Coater, using the Polaron E5550 density thickness FTM unit the sputter coating of the n-PCM samples was performed with nano-layered gold coating.
The PCM and n-PCM samples were inserted in the specimen chamber under vacuum.
The pump pressure was set to slightly over 10 -1 mbar, gold density to 19.3, and the thickness range to 10 nm. The input electrical current was set below 20-15 mA, while the voltage was set to 2 kV before pressed start on the FTM box. The thickness was displayed and when the required value was reached, the process was stopped using the stop button on the FTM box. The gold-sputtered n-PCM sample is shown in Fig. 81.

Imaging
The imaging process followed is described here [320]:

(ii) Selecting a Vacuum Mode
The low vacuum mode was selected for the PCM samples and high vacuum mode for the nanomaterials due to the ease of imaging [333].

(iii) Preparing the Specimen
The sample preparation process was similar as above where in the specimen (sample holder holding the stub with the sample) was inserted into the sample stage and using the ChamberScope and detector was inserted following this to ascertain that it will not touch the detector. The height was adjusted as a 23 mm of WD.

(iv) Inserting the Samples
The Z-stage control was set to the exchange position and the chamber door was held tightly till the vacuum system started pumping. When the vacuum gauge displayed 9 x 10 -5 Torr, the system was considered ready for use.

(v) Focussing the sample area and Image capturing
Using the focus knob, many hit and miss iterations for the optimal focus were required to produce the best possible imaging. For removing stretching effect in the images, astigmatism correction process was used as described in section 5.

XRD analysis of PCM/Nano-material
X-ray diffraction is used to determine the composition of the constituents in a sample semi-quantitatively by comparing the intensities of the diffraction peaks from the known phases [334]. Further, for complex mixtures with more than two phases (even if one amorphous), XRD can still be useful for determining the relative amount of each phase.
The chemical composition of the PCM GR 42 and the purity of 60 nm nano-CuO were tested using XRD technology, based on crystallographic study of the material. Both these materials were outsourced and the manufacturers for PCM were unable to comment on the composition or the reactivity with PCM or nanomaterial. XRD proved a useful tool for identifying the constituents. Also, as it is considered as a good practice to verify the purity of the nanomaterial, XRD analysis of nano-CuO was performed, although manufacturer supplied data suggested > 99 % purity. Siemens D5000 Powder Diffractometer was set to work on a 40 kV, 30 mA, and 50 Hz electrical signals. The data output was in the form of spectra with identified phases representing minerals in the sample. The process for using the XRD is described below.

Sample preparation
As nano-CuO was supplied as 60 nm average particle size, it was used as it was while GR42 as uneven coarse material ( Fig. 83 (a)) was grinded using a tungsten carbide ring mill. Using Normal Mount method, approximately 5 g of powdered sample was loaded into the XRD holder using a clean spatula to spread it evenly. The sample holder ( Fig. 83 (b)) was numbered to ensure the location in the machine was tallied with the sample number. The powder was packed down while avoiding orientation using the narrow edge of the spatula. A glass plate was used to smoothen the powder surface until it became completely flattened and levelled with the top of the sample well (Fig. 83 (c)).

Software program
The XRD commander program was opened and the default location to save the data files was added to the program before adding the input job title and the sample identification details using the Create jobs icon at the top of the screen.

Data interpretation and Analysis
The scanned file was opened on EVA software, and the profile was loaded. Using the search/match toolbar, the following path was followed favour simple patterns mineral, to check the yellow quality marks and the structure selected. Then, also on the main database displaying master, search utility was used to process the peak-searching for matching the minerals in the samples and generating the list of probably matching mineral data. On the toolbar, the pattern tab displayed a list of minerals, from which each mineral was picked one by one to see if their peaks match on the sample profile. Non-matching samples were deleted and the steps were repeated to identify other peaks. The disturbances in the signals or noise in the profiles were removed by replacing the background and the analysed profile was saved as an EVA (or raw data) file.

Conclusions
While the manufacturing and fabrication processes have been discussed in the previous

Introduction
Extensive indoor characterisation was undertaken for the fabricated BICPV-PCM modules, using a range of incident solar radiation intensities supplied by the solar simulator. The electrical performance for the modules such as I-V and power curves were measured and compared. The investigated systems are discussed below: (i) LACPC based BICPV system (ii) SEH based WICPV system (iii) Micro-finned thermal system (iv) SEH based BICPV system The systems were characterised for thermal behaviour in the horizontal direction across the back plate in cases (i), (ii) and (iv) while in vertical direction in (iii) to explore the distribution of heat across the system. The results were analysed both qualitatively and quantitatively. In all cases PCM and n-PCM limited the temperature rise in BICPV to an appreciable degree. Other notable effects with PCM usage such as reduction in hot spot formation and more uniform temperature distribution across the back plate were discovered. The fundamental analytical model available in the literature for a PV-PCM system was challenged due to the detected incongruences in its dimensional analysis.
Therefore, a revised model has been proposed for the BICPV-PCM system. Additionally, a brief cost analysis model for the system have been presented.
The opportunity of integrating a passive micro-cooling system with the back plate of a BICPV module was also investigated. After an initial investigation in case (iii), microfins were found to be beneficial in terms of heat transfer as well as reduced material usage.
Experimental set-up was validated using simulation and uncertainty analysis has been reported for this case. The results from these original experimental investigations were analysed to understand their suitability on a real-world level. This chapter also presents the results from the evaluation of the thermophysical properties of various PCM and n-PCM, such as melting point (onset, peak and end), congealing/solidification point (onset, peak and end), Latent heat capacity, and Specific heat capacity using DSC. A list of different PCM types were analysed since they are envisaged as being useful for related future works. An attempt was made to study the morphological features such as shape and structure of the outsourced nanomaterial and n-PCM using SEM, the micrographs from which are added to this chapter. The particle purity based on material composition was assessed using XRD technology for granulated PCM, GR 42 and the metal oxide based TCE nanomaterial (nano-CuO) has been expressed as diffractograms.

Characterisation of BICPV-PCM system (i) Electrical Characterisation
The short-circuit current (ISC) and the open-circuit voltage (VOC) are the two fundamental parameters of the I-V curve for a PV module. Both ISC and VOC are dependent on the incident irradiance and module temperature in such a way that change in ISC and irradiance are almost directly proportional while change in VOC is only nominally dependent. On the contrary, VOC is inversely proportional to the module temperature causing significant reduction in electrical power at higher temperatures even though ISC increases slightly with temperature [335]. The Pm and η are dependent on the ISC and VOC.
The following sub-sections detail the results derived from the I-V trace obtained for different BICPV modules with and without PCM.

(ii) Thermal Characterisation
The main aim of the work was to improve thermal management of BICPV, the results from the thermal characterisation are presented in the sub-sections below to display the behaviour of the system under different irradiances. The four different systems were characterised first without any cooling media and then passively using PCM, n-PCM or a combination of them with the micro-fins. In one case, tests were performed using water as a passive cooling means to compare the relative effectiveness of the latent and sensible media. In all systems, the thermocouples were attached at various locations under the back-plate and within the PCM containment. The curves represent the temperature obtained (in °C ) with respect to the time elapsed since illumination began.

LACPC BICPV-PCM system (i) Electrical Characterisation:
The LACPC based BICPV-PCM module was illuminated under the solar simulator at  (Fig. 84 (c)). Overall, the average Pm without PCM was 581.7 mW while with PCM, it was 626.4 mW showing a relative efficiency increase of 7.68 %. The absolute electrical efficiency for the module without PCM was 6.48 % whereas with the use of PCM in the system, it increased to 6.98 %. In Fig. 84 (d), a comparison has been made between the percentage changes in Pm at a frequency interval of 30 min, negative values indicates a decline over a period of time. As can be seen, during the first interval, the BICPV experienced an 18.6 % (10.0 % with PCM) electrical power loss. Similarly, during the second interval, the output power reduced by 4.6 % (1.2 % with PCM). In both the cases, power loss after 60 min was nominal, especially with the use of PCM. As can be observed from Table 24, the minimum output Pm using PCM was almost 17 % higher than that without PCM.  The maximum value of Pm, was however higher for the non-PCM case (by 5 %) due to the lower ambient temperature at the start-up. There was a nominal decrease of less than 1 % and 2 % respectively in the minimum and the maximum ISC. Experimentally, it was verified that an increase in module temperature leads to decrement in (Pm) within the aluminium (thermal conductor) backed module.

Effect of irradiance
The effect of increased irradiance on the electrical output parameters and consequently the effectiveness of PCM for the LACPC based BICPV-PCM system was the focus of this study. The experiment as described in the previous section was repeated for 120 min Without PCM @500 Without PCM @750 Without PCM @1200 With PCM @500 WithPCM @750 WithPCM @1200 The results indicated that this particular PCM had a more pronounced effect for lower to medium intensities and was comparatively less effective for higher levels of irradiance which may be due to lower melting temperature range of the PCM corresponding to heat generation in the BICPV panel at those intensities.

Effect of melting temperature range of PCM and insulation
To study the effect of melting temperature range, another PCM, RT55 was used in the set up.

Irradiance (Wm -2 )
Voc without PCM Voc with PCM with lower melting temperature range, RT42). In addition, to examine the negative effects of inappropriate ventilation, the BICPV back plate was insulated with a glass wool of less than 1 cm thickness. The insulated system showed an average Pm of 611.7 mW, reducing the power output by 2.3 %. The profiles for Pm using different backing materials with the module is compared in (Fig. 86). As expected, the absence of thermal regulation (presence of insulation) proved detrimental to the efficiency of BICPV system, in this case by a small percentage for a small period of time. Figure 86: Pm profile at 1000 Wm -2 with and without PCM RT55 and insulated back.

Effect of thermal conductivity
Following the studies on the effect of thermal insulation, it was envisaged to study the benefits of enhancing the thermal conductivity of PCM. The simplest way to achieve this initially was by introducing a metallic (copper) mesh. The results obtained by comparing the maximum power output profiles at 1000 Wm -2 are presented in Fig. 87.

(ii) Thermal Characterisation:
In the LACPC based BICPV system, the temperature at the centre of the module (Tc) was recorded under the plate at an interval of 5 min for four levels of irradiances; 500 Wm -2 , 750 Wm -2 , 1000 Wm -2 , and 1200 Wm -2 .The temperature profiles for Tc and the average temperature at various irradiances is shown in Fig. 89

SEH system with copper tubes (i) Electrical Characterisation:
The experimental results presented here are based on the four test cases of: (a) without applied cooling/with natural convection as the baseline case, with latent cooling means,  6  12  18  24  30  36  42  48  54  60  66  72  78  84  90  96  102  108  114  120  126  132  138   As can be observed from Table 25, the average values of electrical parameters increased with the use of PCM RT50 at both levels of irradiances. The unexpected increase in ISC was possibly due to averaging of the data. The results demonstrated an average increase of 11.5 % in the output electrical power at 1000 Wm -2 and 5.2 % at 500 Wm -2 with RT50 incorporated as compared to a naturally ventilated system for 150 min. Following on from this, the next aim was to study the relative effectiveness of various cooling means at the standard irradiance of 1000 Wm -2 . To achieve this, the profile for the output power Pm (Fig. 92 (a)) and its linear trend ( Fig. 92 (b)) for the WICPV-PCM system were compared with the three cooling means versus natural ventilation. The steady state change (increase or decrease) is shown by this linear trend. Although similar trends of decay in Pm were observed for the system in all cases; the decline was the steepest for the naturally ventilated system followed by the PCMs and it was the least for water cooling. With a higher melting temperature range, PCM RT50 exhibited higher effectiveness and the difference in the trend line compared to the uncooled system showed an increasing gap. As the melting temperature range of RT28 HC was lower than that of RT50, all the available material in the tubes melted earlier and the decline in the Pm started sooner. This indicate that for such systems, a PCM with a higher melting temperature range should be selected; roughly at least 25 °C above the room temperature. The average Pm for the duration of the experiment is shown in Fig. 92 (c). The average electrical power output with no applied cooling was 1.70 W, with RT28HC; 1.97 W, with RT50; 2.00 W and with circulating water cooling reached 2.11 W. The absolute electrical efficiency of the WICPV without additional cooling, with latent and sensible cooling is illustrated in Fig. 92 (d) and one can clearly see the gradual increase from 5.9 % with no cooling to 6.8 %, 6. 9 %, and 7.3 % with PCM RT28HC, RT50 and water respectively. To quantify the relative effectiveness of each of these cooling options, the percentage change in the output electrical parameters of the WICPV with various applied cooling No cooling and other cooling options means compared to no cooling were calculated at 1000 Wm -2 as demonstrated in Table   26. As the I-V data for the run with water was recorded for 90 min, the values from other cooling options were compared for the same duration. The output power was maximum with circulated water cooling (7.8 % increase in ISC, 5.4 % in VOC) followed by using  While the results demonstrated an average increase in the output electrical efficiency of over 17 % at 1000 Wm -2 with RT50, it halved to 6.5 % at 500 Wm -2 as compared to a naturally ventilated system, considering the first 90 min of the experimental data.
Water cooling produced higher improvements in spite of lower heat capacity due to continuous circulation effect. With Sensible cooling with water 6.4 % and 8.0 % higher electrical efficiencies were recorded than RT28HC and RT50 respectively. Water was circulated using water supply tank head as the driving force instead of electrical means, saving electrical expenses. However, in a real world scenario, power losses in running electric motor or pump for water circulation have to be taken into consideration, which may entail active power parasitic losses as well as a higher capital cost. In those cases, the overall efficiency with water systems may even become reduced. Hence, it could be concluded that PCM could still prove to be largely more effective in the long term.

(ii) Thermal Characterisation:
For the SEH system with six copper tubes under the six strings of solar cells, the temperature was recorded at the centre of and under each aluminium strip mounting solar cells. The effectiveness of latent passive cooling media, namely two PCM types; the lower melting RT28 HC and the higher melting RT50 was compared with the sensible passive medium, water, and was expressed in terms of the temperature reduction achieved by them at 1000 Wm -2 . A comparison of the obtained temperature profile for individual channels (channel 1 to 6, expressed as T1 to T6) is shown in Fig. 93 (a-f). As channel T1 was located on one edge of the panel experiencing maximum natural convection, the highest temperature (56.8 °C ) reached there was less than 60 °C . Similarly for T6, which was located on the other edge but experienced slightly lesser natural convection due to its placement towards the solar simulator, the highest temperature (61. It was observed that towards the end, all the PCM in the channels melted to liquid form. The absolute temperature data at the end of the run (at 90 min), within channels T1 to T6 with PCM RT28HC, RT50 and water are arranged in Table 27. As can be seen, both PCM were in molten states as the end temperatures were well above the melting peak for each PCM; 28 °C and 49 °C respectively.  The temperature data for the channels with all cooling options was averaged at 1000 Wm -2 and is shown in Fig. 94. It can be seen that for all channels the application of both sensible and latent cooling were beneficial, though in some channels, one PCM was slightly more effective than another, the reasons for which are unexplored at this stage. The average temperature distribution within channels T1 to T6 for lower and higher irradiances with no PCM versus with PCM RT50 is demonstrated in Fig. 95. One can observe that RT50 was more effective at lower irradiance as the difference between the two graphs was higher. At 500 Wm -2 , the maximum average temperature without PCM   There was only less than a degree of variation between the average highest and lowest temperatures obtained in the case of using both the PCM. It can also be seen that the average temperatures across all channels with water as a cooling medium was almost constant with a variation of ±0.4 °C. As an exception, T1 (inlet water channel) experienced the lowest average temperature of 37.5 °C in contrast to the highest 38.6 °C within T2 (outlet channel) due to the water circuit configuration. For the two cases of PCM cooling, the temperature at the central channels (T3 and T4) were comparatively higher due to their central location achieving lesser cooling via natural convection otherwise available to the channels on the sides or edges such as T1. The average temperature reduction achieved across channels 1 to 6 using different passive cooling media as compared to the case with no cooling for the duration of the experiment is highlighted in Table 29. As expected, the circulated water achieved the highest temperature reduction across all channels followed by RT28HC and RT50 respectively.  respectively. This indicates that PCM circulation could give better results though it may incur additional costs in pumping and transportation. It may be worthwhile to highlight that for the experiments on SEH based WICPV system, water was circulated using the head of the water supply tank as the driving force instead of electrical means, thereby saving electrical expenses.
The use of PCM increased the uniformity in temperature distribution throughout the module, which is an important outcome for BICPV systems as a thermal disequilibrium leads to the formation of hot spots. The temperature distribution throughout the module was studied and it was found that the average temperature difference between the channels T3 and T4 (located in the centre of the module) and T1 (on the edge) was higher

SEH-n-PCM system (i) Electrical Characterisation:
The SEH based BICPV system was characterised at a lower (500 Wm -2 ) and a higher (1000 Wm -2 ) irradiance for a duration of 150 min at 0° inclination. The system was then  Fig. 97 (a) at 500 Wm -2 and (b) at 1000 Wm -2 . As can be seen, at both lower and higher intensities, the use of n-PCM 1 exhibited higher improvement in performance for the system as compared to n-PCM 2.
Though with n-PCM 1, VOC was highest at lower irradiance, it was highest with n-PCM 2 at higher irradiance, although the disproportionate decrease in ISC caused a decline in the overall Pm. As shown in Table 30  From the experimental data, it can be concluded that n-PCM 1 was more effective, with an increased electrical power output at both irradiance levels and n-PCM 2 performed relatively better at lower irradiance level of 500 Wm -2 .
(ii) Thermal Characterisation: The SEH based BICPV system was thermally characterised without PCM ( RT 42) cooling first and then the thermal performance was compared with n-PCM 1 and n-PCM There have been minor fluctuations in the observation for the curves due to external environmental changes beyond control such as occurrence of day-night temperature difference as well as different temperatures (weather dependent) every day although every precaution was taken to minimise these effects and ensure that the working /operating conditions remained almost identical for all test cases. Apart from minor fluctuations, there was a common trend followed with the highest temperature monitored at the centre of the module. For comparing the temperature profiles with no-cooling, n-PCM 1 and n-PCM 2 at significant points under the BICPV back-plate for 1000 Wm -2 and 500 Wm -2 , Fig. 99 and Fig. 100 can be referred to, respectively. At most of the monitored points, n-PCM 1 proved to reduce the temperature better than or at least equal to as n-PCM 2, with a minor exception at 1000 Wm -2 (T7).      T1  T2  T3  T4  T5  T6  T7  T8  T9 T1  T2  T3  T4  T5  T6  T7  T8  T9   During the experimental run, the temperatures across the system were recorded at these points on a vertical section: (a) T1 or ambient temperature above the insulation layer, (b) T2 above the plate centre, (c) T3 below the plate centre, and (d) T4 at the centre of the PCM containment. A comparison of the average temperatures obtained across all the six configurations are given in Fig. 102. The bar graphs were plotted in such a way that the effect of introducing micro-fins on the un-finned surface can be easily compared for the three cases of using natural convection, PCM and n-PCM. As can be observed, the temperatures T2, T3 and T4 were recorded highest for the un-finned configuration, followed by the micro-finned configuration with only natural convection cooling. The same trend was followed for the configurations entailing the use of PCM and n-PCM respectively. It is to be noted that the ambient temperature closest to the shielded set-up,    The effect of nanomaterial addition seemed more pronounced with the un-finned surface as compared to the micro-finned surface because the difference between the two curves was greater than for un-finned surfaces. The reason behind this is unexplored so far, and may require further detailed analysis to confirm.  As can be noticed, the micro-finned systems exhibited more effective thermal regulation as compared to the un-finned systems. In addition, the configurations using PCM demonstrated higher temperature reduction in comparison to non-PCM systems and consequently n-PCM performed better than PCM in thermal regulation. Another interesting theoretical aspect of this attained temperature reduction can be its correlation with the increase in the output electrical power. Based on an analytical method, the achieved temperature reduction could be converted to the increase in electrical efficiency and consequently power production for a BICPV. Assuming the temperature coefficient of maximum power point as -0.5 %/°C [52], which is a ball parked value for the BICPV panel, 12.5 °C decrease in its average operating temperature using n-PCM with microfins will theoretically lead to an increase of as high as 6.25 % in output power. Similarly, with the use of PCM and micro-fins, an 11.2 °C reduction in average temperature could increase the maximum output power by 5.6 %. Even in worst case scenarios of either using PCM with un-finned plate or PCM with micro-fins, a reduction of 10.7 °C and 9.6 °C will also result in 5.35 % and 4.80 % of electrical power gain respectively. It is worth noting the effect of each configuration on the heat transfer (Fig. 106). Indeed, both for the un-finned and the finned surfaces, replacing air with PCM leads to an increase in heat transfer coefficient between 16 % and 35 %. These results further confirm the advantages of using PCM in heat transfer applications and its potential as an effective coolant in BICPV systems. Figure 106: Heat transfer coefficients registered for the different configurations, with error bars. On the right axis, the overall fin effectiveness of each configuration compared to the un-finned surface exposed to air is reported. By definition, the effectiveness of the un-finned configuration is equal to 1.
On the other hand, in accordance with the previous literature [124], the introduction of fins causes a reduction in the heat transfer coefficient. A lower heat transfer coefficient should not be implied as a cause for reduction in thermal performance since it is mainly due to the fact that the heat transfer coefficient is calculated on the total surface area, which is larger when fins were added. Indeed, by analysing the overall fin effectiveness, defined as the ratio of the heat transfer rate for each configuration to that of the natural convection based un-finned case, it can be seen, that a combination of micro-fins and PCM can lead to an enhancement as high as 32 %. The thermal performance of the unfinned plate under the naturally convective condition was found to be superseded by the one with micro-fins, PCM, and n-PCM in increasing orders of effectiveness.  (Table 32). As a result, RT42 was 9.7 % more effective than GR42 in reducing the maximum temperature and by 13.2 % in reducing the average temperature.  Various combinations of mass fractions can be used to produce a PCM mixture exploiting the non-leaking nature of GR42 while high utilising the higher heat capacity of RT42.
The ideal fraction would contain GR42 saturated with RT42, but at which there is no visible leakage upon melting of the PCM.

DSC Analysis
The data obtained for two samples were prepared from the bulk material and three runs were performed on each sample. The thermograms obtained for a PCM sample with respect to Sapphire (standard material, shown by baseline) are given in Fig. 108 as an example. The three curves with different colours denote the different runs using the same samples. For pure materials, only single peak is expected to appear, however, the other smaller peaks in PCM appear as a result of polymeric nature of the material with different combinations of hydrocarbons present in it.
In order to interpret the DSC curves correctly, it is imperative to examine it for obvious artefacts, which are not caused by the samples, to avoid any misinterpretation of the results. Appendix-1 (Fig. 116) shows the examples of artefacts widely accepted [327].
The results for melting point (peak, range for onset and end), congealing point (peak, range for onset and end), specific and latent heat capacities for the many tested PCM and n-PCM samples are highlighted in Appendix-3 (Table 38-

SEM analysis
The micro-structural analysis of, nano-CuO and the resulting n-PCM using SEM has been presented in this section. The morphological characterisation was carried with the aims to evaluate the degree of dispersion of the nanomaterial in the n-PCM after the calorimetry cycle test. The purpose of carrying this exercise out was to assess if the structure of the material provided by manufacturers before shipping differed from them just before use, which may indicate any significant differences within the thermo-physical properties. As mentioned in the previous chapter, the quality of images for the n-PCM samples did not produce satisfactory results with the first SEM run, so a different route of gold sputter coating the samples was chosen before scanning using second SEM run.
The images were still not of highly acceptable quality as paraffin wax samples started to charge (white appearance in dark background). The charging effect was observed due to their non-conducting nature. SEM imaging was performed for the nanomaterial CuO (60 nm) and an n-PCM (RT42 enriched with 1.0 % CuO).

Morphology of CuO
The surface morphology of the nano-CuO sample supplied by the manufacturer on a 1 micrometer (μm) scale (Fig. 121, Appendix-4). . It was cross-validated against the approximate particle size of a prepared layer of the nano-CuO sample. The nano-CuO micrographs are shown on different scales and resolutions in Fig. 122 (Appendix-4). The island growth of the tightly packed spherical arrangement can be clearly seen, the diameter of which varies between 65 and 150 nm. Although the supplied material sample was ground, milled and dispersed using solvents such as ethanol, agglomeration of nanomaterial can be clearly observed, showing a non-uniform distribution of the particle size, heterogeneous morphology and presence of other material.

Morphology of n-PCM
The n-PCM samples were first imaged without any metallic sputter coating on a scale of 200 μm and 10 μm respectively. However, to avoid the accumulation or build-up of static electric charges on the specimen surface that appear due to non-conducting nature of paraffin waxes, the samples were coated using gold sputtering. Appendix-4 shows the micrographs obtained on gold coated samples on a scale of (c) 400 μm, (d) 200 μm, (e) 20 μm and (f) 6 μm scales. One can clearly see from the figures that the charging of PCM continued to happen (bright patches seen in Fig. 123 (a) and (d)). The clarity of the images decreased with the increasing resolution and the dispersion of nanomaterial in the n-PCM could not be experimentally determined with the micrographs.

XRD analysis
Using X-Ray Diffraction, the samples were quantitatively analysed to determine the phase, structure and crystallinity. This was done by comparing the intensity of the diffraction peaks from each of the constituents to the intensity of the known phases of the pure material. Structural properties and crystalline structure of materials were also investigated using XRD technique. The rationale behind this exercise was that: (a) the nano-CuO sample didn't yield sufficiently satisfying results using SEM, possibly due to agglomeration, hence the supplier provided chemical properties had to be cross verified for the purity of nano CuO and (b) the composition of granulated PCM, GR42 wasn't known from the manufacturer's data. Further, they were unable to comment on the reactivity of its constituents upon treatment with other PCM or nanomaterial. The results from XRD analysis are described in this section.

Analysis of CuO
The XRD spectrum of nano-CuO nanoparticles provided by the manufacturer (though not very clear) is shown in Fig. 109 and it can be seen that the there are other constituents in a larger proportion in addition to some other materials in smaller ratios. The diffractogram image obtained experimentally is given in Fig. 110. Two reflections were observed in the diffraction patterns at 2θ = 35.6 and 2θ = 38.8 similar to [337], which attributes to the formation of the CuO as a monoclinic crystal. It was beyond the scope of the XRD programme to derive the weight fraction of the constituents in the samples. However, the presence of any common mineral or metal was easily gathered with the help of the analysis software. It was found from the diffractogram that the sample contained Tenorite (CuO), as well as Cuprite (Cu2O) forms of copper oxides in addition to metallic copper (Cu). Tenorite occurs in the weathered or oxidized zone associated with deeper primary copper sulphide orebodies and Cuprite is an oxide mineral and a minor ore of copper, which can easily oxidise into other forms. As can be seen, there were some percentage of Cu2O and metallic Cu present in the apparently pure sample of nano-CuO (inset of Fig.   110). It can be deduced that approximately 75 % of the material sample comprised of CuO while the rest was 17 % Cu2O and 8 % metallic Cu. Figure 109: Manufacturer provided XRD image of the 60 nm Nano-CuO (Image was provided with an intrinsically low resolution, which can't be altered).

Analysis of GR42
For the granulated sample of PCM GR42, Fig. 111 displays the XRD spectrum. From the data it was found that the sample was composed of Paraffin wax in addition to Quartz and Haematite, all of which were non-reactive to other paraffin waxes as well as nanomaterials. The exact percentage composition wasn't required as this task was performed to detect the presence of any reactive constituents.

Experimental uncertainty and setup validation
The uncertainty of the experimental and the calculation processes was measured as described by the previous works in this field [124], by applying the formula on propagation of errors for independent variables [338]. In particular, the error occurring in determining the heat transfer coefficient of air in an un-finned configuration was calculated to determine the experimental uncertainty, where the heat transfer coefficient, h, is defined in Eqn. 21 [75]: Where Qin is the input power, Afins is the finned surface extension and Tfins and Tair are the steady state temperatures of the fins and of the air inside the containment respectively.
Therefore, the uncertainty (Uh) was determined in Eqn. 22 [75]: Where the prefix U indicates the absolute uncertainty for each variable.
In accordance with [123], a conservative uncertainty of ±4 % was considered for the surface area of the micro-finned/un-finned plate. The applied electrical power was calculated from the supplied I, V measurements, which were measured in the proximity of the input connections to the heater. In accordance with the previous studies, no voltage drop was considered to have taken place in the wires between the power supply and the heater. All the uncertainties used in this analysis are shown in Table 33. Furthermore, an average deviation of 1.0 % was measured between the voltage readings and thus, taken into account for this calculation.  [123] and is below the 20 % uncertainty span generally considered acceptable for natural convection correlations [339].
The quality of the experimental setup was then validated through a 2D model built in COMSOL Multiphysics® 5.2. Considering the simplest case for an initial investigation into validation modelling, it was based on a sufficiently appropriate 2D geometry. The behaviour of the flat surface when exposed to air within the containment, was modelled and then compared with the experimental results. The flat aluminium plate was reproduced on top of the setup (Fig. 112 (a)), the containment box ( Fig. 112 (b)) was modelled using the built-in properties of acrylic plastic (1470 J/kg.K, 1190 kgm -3 and 0.18 Wm -2 K) and the centre of set-up is depicted in Fig. 112 (c)

Other Important Observations
During the course of the experiments, a few important factors were noted which may either have internally affected the experimental outcomes to a minor extent or have appeared as an outcome of the experiments. These are listed and described below.

White spots/patches
Undergoing illumination for extended periods of time, white spots or patches between the encapsulant layer and the solar cells started to appear in the BICPV modules. Fig. 113 shows the white patches appearing between the soldered LACPC cells and Sylgard layer

T1
after approximately 20 hours under the solar simulator at an intensity of 500-1200 Wm -2 .
The effect of these could be sensed with the decrease in the output short circuit current from the module. There hasn't been much study on these patches, however, the authors of [340], who have used analogous systems for their research, described similar observations. They indicated that it may be due to uncured silicone material that gives rise to the white spots as the system is exposed to higher temperatures for elongated period of times. In addition, presence of certain materials or chemical interactions may inhibit complete curing of Sylgard such as the flux solvent used for soldering the solar cell with tabbing wire.

Nanomaterial agglomeration
The n-PCM for the micro-finned systems was synthesised by mixing 0.5 % (by mass) nano-CuO with PCM RT42 at 60 °C to ensure the fully melted state of the PCM. The mixture was then ultra-sonicated using a Hilsonic® ultra-sonicator machine for 24 hours.
The possibility of agglomeration within the nano-PCM was minimised by using the ultrasonic vibrator [250]. A low mass fraction of CuO was selected for the initial testing as the lower concentrations of TCE were found to show higher energy storage capacity with lesser costs involved [284] . Another important observation during the course of the experiment was the limitation of mixing nano-CuO with paraffin wax; namely the segregation, agglomeration and deposition of the metal oxide nanoparticles due to the difference in their densities (Fig. 114). After successive heating and melting cycles, the n-PCM showed visual signs of agglomeration and deposition of nano-CuO due to the difference in their densities. In future work, other forms of TCE or nano-particles with a density of the similar order as the base PCM could be used to overcome this. Though some tests have been performed with higher mass fraction of 1.0% for BICPV system and DSC, it was envisaged to investigate the detailed effect of variation in mass fraction on performance in a future work, and hence this consideration does not fall in the immediate scope of this thesis.

Leakage Analysis
Leakage is one the most challenging practical issues [257,258]

3-D printing
The PCM containments for latter studies were fabricated using 3-D print technology. This method not only contributed to substantial leakage control, but also reduced the high manufacturing turnaround time, and the associated labour plus material costs. This approach for manufacturing PCM containments using additive layer manufacturing technology has not been reported elsewhere and could pave the way for its future use.

Cost-Benefit Analysis
In order to analyse the cost effectiveness of the individual elements (PCM, n-PCM and micro-fins) forming the passively cooled system in tandem, a simple excel based costbenefit model was produced. It compared the additional costs per degree temperature reduction (£/°C) for the system both individually and combined and was based on the amount of material used. The results from the cost-benefit analysis is shown in Table 3 and the details are given in Appendix-2. The assumptions are stated below: 1. It was assumed that the amount of base PCM did not change in n-PCM and that nanomaterial was just an add-on as the mass of the nanomaterial was negligible.
2. The costs of micro-finning per plate has been assumed as £ 2.0 as a ball-parked value.
3. The material costs are the actual costs per gram of the material used.
In this analysis the following were not considered: 1. The cost of 3-D printer and manual labour costs in designing the component.
2. The material cost saving due to micro-finning the surface instead of using thicker material section for producing the BICPV back plate.
3. The costs of micro-milling machine and other tools and accessories such as glues etc. As can be observed from the Table 34, the cost of using n-PCM for every degree of temperature reduction was slightly lesser than the cost of using PCM. This was due to higher temperature reduction achieved with n-PCM although the overall cost of the n-PCM was relatively higher by 45.4 % compared to the PCM costs. The cost of having micro-fins only was the highest for temperature reduction and the cost of micro-fins with n-PCM followed the one with micro-fins with PCM..

Module efficiency degradation
It was observed that the BICPV modules lost a part of their absolute electrical efficiencies over a period of time as they were constantly under intense irradiance levels during indoor characterisation. Though this constant high irradiance (750-1200 Wm-2) will definitely not be the case in real outdoor conditions, the lifespan of the modules must still be higher.
However, with the analysis, it was found that on an average a module was degraded by

Discussions
Going back to the research questions outlined initially (section 1.7), this section aims to describe whether and how have they been addressed. While the qualitative assessment has been detailed below, quantitative results are presented in the next section.
i. What will be the quantified electrical and thermal performance enhancement In all the four BICPV systems, addition of PCM achieved at least a 7 % increase in the output average electrical power. The heat absorption in PCM also reduced the overall average temperature of the BICPV panel. It may be concluded that these original experimental investigations to establish the PCM effectiveness for thermal management of BICPV were successful, a field which was unexplored before.
ii. Which combination materials can be synthesised via enriching PCM with nanomaterials? What will be the morphological and thermo-physical properties?
To address this, a low cost, relatively easily available, and non-toxic nanomaterial, nano CuO with an average particle size of 60 nm was dispersed in PCM RT42, in two mass ratios of 0.5 % (n-PCM 1) and 1.0 % (n-PCM 2). The morphology of nanomaterial was studied using a scanning electron microscope but it did not match with the manufacturer provided data for particle size. The reason for this was the agglomeration of the particles to form bigger lumps at a nanoscale. However, this did not affect the experimental results vastly on a macroscopic level. The dispersion of the n-PCM within PCM was also attempted to be studied using SEM, but could not yield meaningful results due to the charging of the PCM. The alternate solution, to sputter coat the sample with gold nanolayer, also did not prove effective. The inevitability of this issue was rooted in the nonconducting nature of the paraffin waxes. Nevertheless, the obtained images and micrographs are added within this thesis. The thermal conductivity enhancement provided by adding low proportions of nano-CuO (with k =33 W/m·K), was very little as the effective thermal conductivity of the n-PCM was calculated to have a 0.35 % increase.
This could be increased with materials with higher thermal conductivity such as carbon nanotubes (with k of the order of 2000 W/m·K).
iii. What will be the expected performance enhancement realised within the mentioned BICPV system using laboratory synthesised n-PCM?
With respect to PCM, n-PCM proved effective by over 3 % in terms of improving electrical power output in one BICPV system. The n-PCM 1 proved more effective for temperature regulation in two experimental cases compared to n-PCM 2.
iv. How does variation in the following parameters affect the output power from, and the efficiency of, the BICPV in question?
 Levels of concentration -In order to study the effect of these parameters, various concentrators such as LACPC, SEH with different concentration ratios, were used for fabricating the BICPV and no major differences were observed with regards to PCM application.
 Solar irradiance levels -A range of irradiance (500 Wm -2 , 750 Wm -2 , 1000 Wm -2 , and 1200 Wm -2 ) were selected to repeat the experiments. It was found that PCM RT42 was more effective at higher intensities of over 750 Wm -2 , due to its melting range corresponding to the temperature rise in the BICPV.
 Heatsink designs -The heatsink design for PCM containments in all cases were based on a previously reported analytical model for flat-plate BIPV-PCM systems from literature. However, due to mathematical inadequacy, a new and refined BICPV-PCM model for designing was proposed.
 Material used for PCM containment -The two materials tested for fabricating the PCM containments were 13-mm thick Perspex sheet (cast Poly-methyl methacrylate) and 3-mm thick ABS plastic (Acrylonitrile butadiene styrene). It was easier to machine the latter and due to its lightness, the overall system bulk of the BICPV-PCM system wasn't increased, an important aspect for such systems. In addition, ABS based PCM containment was 3-D printed as a single part, exhibiting nominal PCM leakage upon melting into liquid form.
 Heat exchanging media -in addition to latent heat media (PCM), a sensible medium (water) was used to compare its effectiveness. Although water proved more effective, it was predominantly due to the circulation effect. although micro-fins alone provided little cooling potential compared to the un-finned surface. Further, micro-finning the back plates were also envisaged to reduce the system bulk and reduce the mass specific power of BICPV systems.
vi. What will be the quantifiable effect of applying a sensible versus a latent heat media, to the thermal regulation of BICPV systems?
As expected, a sensible media has far more potential for temperature control as compared to a latent media. However, the experimental results revealed that the difference in the performance wasn't too great, with only 8 % and 6 % higher electrical efficiencies compared to the two PCMs with a lower and a higher melting temperature range. This analysis did not consider any energy expenditure for the circulation of water since the natural supply head was the driver. If this was included, potentially the system may hardly reach a break-even for the cooling with water scenario or it may prove even more expensive or power consuming than the additional power produced

Conclusions
The detailed indoor experimental characterisation of the fabricated BICPV-PCM systems for electrical and thermal performances were performed using a highly collimated solar simulator at selected solar radiation intensities, and at a 0° angle of incidence.  Using XRD analysis, GR42 was found to be composed of Paraffin wax, Quartz and Haematite, and nano-CuO sample contained Cuprite and metallic copper.
 Segregation, agglomeration and deposition of the metal oxide nanoparticles was observed in n-PCM due to the difference in their densities.
 White spots were formed in BICPV over a period of continuous use at high irradiances. These were responsible for the degradation of output efficiency.
 The cost per degree temperature reduction using n-PCM was slightly lesser than that using PCM. The cost (£/°C) of having micro-fins only was the highest at 1.54 for a unit temperature reduction and the cost of micro-fins with n-PCM (0. 19) followed the one with micro-fins with PCM (0.23).
From the results obtained within this chapter, it could be concluded that the introduction of passive cooling mechanisms such as PCM, n-PCM and even producing micro-finned back-plate for BICPV systems have been highly effective in thermal management as well improving the electrical efficiencies. PCMs have also proved to improve the uniformity in BICPV average temperatures thereby offering a potential to increasing their lifespan.

Chapter 6. Conclusions and Recommendations
This final chapter draws conclusions from different aspects and phases involved in the experimental work carried out within the thesis; material selection, design, fabrication of components, assembling systems and characterisation. It presents an overall synopsis of the lessons learnt, limitations of the work, the achievements and future perspectives for the research area, with focus on adopting new materials, methods and approaches.

Summary
This thesis conducted experimental investigations on innovative solutions to improve the electrical performance of BICPV using passive temperature regulation. The research

Selection of materials and components
Within this study, the solar receiver/BICPV is considered as the heat source. The PCM system, which acts as a heat sink, consists of PCM containment and organic PCM/n-PCM.
A summary of the most common materials available for BICPV-PCM system and the suitability of the selections have been presented. Among the various available materials, most appropriate materials were selected on the basis of a literature survey.
 Organic PCM were used for latent heat removal as they are safe to use (non-toxic), cheaply available and are approved for building use in many countries.
 Finer nanomaterials with smaller average particle size (60 nm) were more efficient with prolonged dispersion in base PCM, increasing the effectiveness of the n-PCM in the long run.
 Despite adding nominally to the costs of machining, micro-finned plate was used to assess its effectiveness in possible applications as a BICPV back plate.
 In the first instance of testing the effects of k enhancement on PCM effectiveness, a copper mesh was selected due to low costs and availability.
 Nano CuO material was preferred due to higher thermal conductivity although metallic nanomaterials are more conductive, they cost 6-7 times more. In addition, the density of Cu is higher which leads to more particle segregation reducing its effectiveness over continuous charging/discharging cycles.
 Finer nanomaterials with smaller average particle size (60 nm) were more efficient with prolonged dispersion in base PCM, increasing the effectiveness of the n-PCM in the long run.

Conclusion from design and manufacturing aspect
One of each four different BICPV systems were fabricated using LACPC and SEH concentrators with a low concentration ratio (< 10x) as per the standard process in literature. However, the thick glass back-plate was replaced with a thin aluminium plate.
The top glass cover was also removed and PCM containment was added at the back which reduced the bulk of the BICPV by more than 5 times.
 For same output power, the LACPC-BICPV system with a concentration ratio of 2.8 saved almost 50 % of the required number of solar cells and SEH based BI/WICPV required 60 % less PV material compared to a flat PV.
 The PCM containment was first produced by mechanical means and the components were manually picked-and-placed. However, PCM leakage indicated a need for improvement. PCM containments for latter studies were fabricated using 3-D printing, which controlled leakage substantially and in addition, reduced the high manufacturing turnaround time, and material costs.
 Micro-fins proved beneficial in terms of higher mass specific power or mass to power ratio. System bulk is an important consideration in BICPV.

Conclusion from electrical characterisation
Overall, PCM into BIPCV increased Pm by various percentages depending on melting temperature range of PCM and irradiances. The results for electrical characterisation have been concluded in Table 35.
PCM effectiveness is presented in terms of increase in the average output electrical power (Pm). The salient points are listed below:  PCM proved more effective at higher (1000 Wm -2 ) compared to lower intensities (500 Wm -2 ). This was a direct result of PCM melting temperature (38 °C-43 °C).
PCM with lower range may prove effective, but due to 30 °C ambient conditions in the laboratory it was not possible to select a lower melting PCM.
 Introduction of copper mesh as TCE showed a 1.7 % increase in the average Pm.

Conclusion from thermal characterisation
The thermal characterisation of the four different BICPV-PCM systems; (i), (ii), (iii) and (iv) the micro-finned PCM system are presented here. The effect of using PCM in terms of maximum and minimum temperature reduction is given in Table 36 and the average   values are compared in Table 37. It has been observed that PCM/n-PCM have been able to reduce the maximum system temperature between 6 % to 34 % depending on the configuration while the average temperature by at least 7.5 % and a maximum of 30 % compared to the systems with no PCM.  The module centre showed significantly higher temperatures than other parts.
With the use of PCM, temperature variation was reduced for all cases and experimental conditions. It can be inferred that PCM can reduce thermal disequilibrium within the module, reduce hot spots formation and avoid untimely degradation of the modules.
6.6 Material characterisation 6.6.1 DSC  A list of various types of PCMs either used within this thesis or are considered useful for future work were analysed for charging (melting)/ discharging cycles.
At 10 K/min, measured values were substantially higher for almost all PCMs, while at 2 K/min, supplier's data proved closer to measured values.
 The melting onset temperature was slightly decreased at higher heating rates in n-PCM 1 and substantially increased in n-PCM 2. The change in experimental values for PCMs compared with supplier's data were 12 % on either side.
 Nanomaterial addition increased the solidification end temperature for most heating rates compared to the base PCM. The experimental means were lower than manufacturer's values, suggesting a higher degree of super cooling.

XRD
 It was concluded that GR42 composed of Paraffin wax, Quartz and Haematite, all of which were non-reactive to other paraffin waxes as well as nanomaterials.
 The diffractogram confirmed the presence of Cuprite (Cu2O) forms of copper oxides in addition to metallic copper (Cu) in the nano-CuO samples with approximate content of 17 % and 8 % respectively.

Cost-Benefits
 The cost per degree temperature reduction using n-PCM was lesser than PCM due to higher temperature reduction, although the material cost for n-PCM was higher by 45.4 %. The cost (in £/°C) of micro-fins only was the highest at 1.54 while for micro-fins with PCM, it was 0.23 and micro-fins with n-PCM at 0.19.

Limitations
The limitations of this thesis and the experimental work are discussed as follows:  Characterisation: taken place indoors within the lab, so outdoor behaviours of the BICPV and BICPV-PCM systems are not included. The performance deterioration was also apparent due to overheating of cells for prolonged hours. Hence the system did not show high stability over time.
 While casting concentrators, some micro-bubbles were trapped inside the liquid.
This may have affected the clarity as well as the effective concentration ratio.
 The solar cells were individually tested after soldering but a small minority of them failed during experiments, due to their ageing. not considered for the initial phase of TCE investigations.

Achievements
As part of the UK Department of Energy and Climate Change 2014 Solar PV Strategy [4], the emphasis as a key target is changing significantly towards increasing solar PV rooftop installations. The present study was conceived as an incremental contribution towards those goals by helping unlock the huge potential for buildingintegration of PV across the UK. The salient points of achievements are given below:  This thesis reports novel approach for enhancing electrical performance of the present day low-concentration BICPV passively with PCM which has not been reported before. This is the first instance of work relating thermal management of BICPV to efficiency increase. In previous studies on BIPV-PCM, primarily focus was on temperature regulation and melt fraction, i.e. the PCM containment part of the system. In this study, however, thorough investigation was carried on the effect on electrical parameters including efficiency, output ISC, VOC and Pm .
 As yet, the proposed mathematical model for BICPV-PCM systems is the only optimised analytical model for designing such systems. Previously used BIPV-PCM model was mathematically incongruent so, revised model was proposed.
 BICPV cooling using simple passive means such as micro-finned back-plate combined with PCM/n-PCM were developed. The low thermal conductivity of paraffin wax based PCM was successfully addressed using inexpensive TCE.
 The experimental investigation in this study has allowed identifying the most favourable passive cooling solution to BICPV temperature rise. The use of microfins with PCM and n-PCM for the passive cooling of BICPV have not been previously reported. The micro-fins showed dual benefits: reducing the system temperature and material and mass; enhancing the mass specific power.
 The use of PCM and n-PCM has proven to reduce the hot-spots in BICPV; a local overheating phenomenon leading to damaging effects such as failure of solar cells or cracking of encapsulating layer exposing the cells to environment.
 To address practical challenges such as leakage, high manufacturing turnaround time and associated costs, 3-D printing was used to fabricate PCM containments.
This approach for containments manufacturing using additive layer manufacturing has not been found in literature and could pave way for the future use with market penetration.
 The specific knowledge accumulated on the issues and the challenges related to fabrication and long-term working of BICPV-PCM systems have been recorded, contributing to the development of more reliable and competitive systems.

Future Recommendations
The major objectives of the project were achieved and are reported in detail in the present thesis. However, a full scale outdoor characterisation of the BICPV-PCM systems would allow refining the presented design and material. A long term outdoor testing would also prove the durability of the developed BICPV-PCM systems and highlight any modifications that may need introducing. Moreover, BICPV with the micro-finned back plate with PCM and n-PCM in outdoor testing will lead to more accurate results on the PCM effectiveness as well as provide the data for studying their discharging. Finally, new ideas to be investigated have emerged during the current study. Hence, a list of recommendations for future works is reported here. They are categorised into component level, system level and application level:

(i) Component-level recommendations:
 Improve BICPV quality fabrication by automated processes for soldering solar cells and manufacturing concentrators. Use more appropriate soldering wires (material and dimensions) to avoid resistive losses within the internal circuitry.
 Other n-PCM such as CNT nanomaterial, in varying mass proportions, can be used. Inorganic PCM such as SP 31 or powdered PCM such as PX 25 can also be investigated, for which the thermal analysis are presented here.
 Composite PCM (varying mass proportions of granulated/organic PCM) such as GR42+RT42, can be treated with metallic nano-material for ideal combinations of leak-proofing and higher thermal conductivity.
 Finer nanomaterials (60 nm average particle size) helps in more efficient dispersion and for longer periods in the base PCM although CNT would be the best choice as their densities are almost a quarter of the density of nano-CuO.
 Encapsulation of PCM in Aluminium or HDPE shells for improved leakage control, ease of manufacturing, and fire safety of the system.

(ii) System-level recommendations:
 Investigate the relationship between micro-fin parameters (effectiveness, geometry etc.) for optimising heat sinks designs. Benefits for CPV cooling due to introduction of micro-pin fins instead of plate-fins should be studied.
 The stored thermal energy in the PCM can be used for potential building heating application either with thermoelectric devices or heat exchangers with other mediums such as water. The thermal energy extracted from modules can also be utilised for low temperature applications e.g. water and air heating [84].  PCM discharge studies can assess the possible uses of heat and integration with a secondary power generator (Peltier or Seebeck devices).
These recommendations can improve the BICPV performance on one hand and reduce the overall costs of these systems on the other hand, by increasing BICPV lifetime. These investigations could not be conducted within this thesis because of (a) the time constraints, and (b) out of the main scope of PhD.
As a consolidated conclusion, it could be established that the research problem of determining the effectiveness of introducing micro-finned back-plate, PCM and n-PCM into BICPV systems has been answered to a substantial degree with the quantified outcome but further questions are raised concerning the long term stability of these systems. The experimental results presented within this thesis provide promising foundations on which to build the next stage of the work. By introducing micro-finned back-plate, PCM and n-PCM into a utility-scale BICPV system, it would be possible to test their effectiveness not only through temperature regulation, but also in terms of enhancing the output electrical efficiency of the BICPV systems. These types of investigations can contribute immensely to the future of the BICPV market, opening pathways for innovation and market penetration to increase its share in the competitive power generation industry. Figure 116: Artefacts caused by various sources, a-c for exotherms and d-h for endotherms [327]. c: noisy signal caused by an entry of cool air into the cell due to misaligned cell lid d: electrical effects either due to a possible discharge of static electricity in the metallic parts of the system, or disturbance in power supply (d1), or frequency interference caused due to the presence of radio emitters, such as a mobile phone in the vicinity (d2).

Appendix-1
e: sudden change in the ambient temperature due to day-night cycle or presence of heat producing instruments.
f: if the increasing vapour pressure of the sample causes a burst of the lid, it produces an endothermic peak with a height proportional to the quantity of vapour.
g: periodic/intermittent closing of the hole in the lid due to the condensing droplets.
h: contamination of the sensors due to the presence of any residues from previous experiments

Appendix-2
The cost-benefit model of the micro-finned thermal management system for the BICPV. Appendix-3

Melting point (peak)
The melting peak determined as the maximum value of the DSC signal during the melting cycle is shown in Fig. 117 which is an example for RT42 at a heating rate of 5 K/min.

Congealing/Solidification point (peak)
The solidification peak temperature was determined as the maximum negative DSC signal during the solidification process as shown in Fig. 118 as an example for RT42 at 5K/min. Table 39 lists the solidification peak temperatures for the PCMs investigated.  It was evident from the data that nano-CuO addition increased congealing temperature.
In general, for most other PCM, the manufacturer's values were closer to the measured values at the lowest heating rates of 2 K/min. At higher heating rates the measured values were substantially higher for almost all the PCM.

Melting Onset point
The PCM melting temperature range is based on the start and the end of the melting process throughout which the melting of PCM remains in progress. The onset point was determined as the intersection of the starting slope and the horizontal line (Fig. 119).

Solidification onset point
The congealing or solidification temperature onset was determined in a similar way as melting onset, but on the opposite side of the curve. Solidification onset temperatures for different PCM are presented in Table 41. Nanomaterial addition increased the mean onset temperatures, especially at the lower heating rate of 2 K/min.

Melting End point
The PCM melting end point was determined as the intersection of the ending slope and horizontal line (Fig. 120) and the values obtained for the PCMs tested are listed in Table   42. Both the n-PCM exhibited lower melting end point temperatures than the base PCM. Figure 120: Example of melting end temperature assessment: RT42 heated at 5K/min.

Solidification End point
The solidification end point temperatures for the PCM were determined as the intersection of the ending slope and the horizontal line. The data for PCMs is tabulated in Table 43.
For most cases, nanomaterial addition increased the solidification end temperature compared to the base PCM. For the other PCMs, the experimental means were lower than manufacturer's values, indicating the degree of super cooling would be higher during congealing cycle than expected.

Latent heat capacity
The latent heat capacity was calculated as the area under the Cp curve and the temperature line and was determined on the basis of the highest peak area values. The data for melting and solidification cycles are in Table 44 and 45 resp. It can be observed that for n-PCM 1, the mean values were lower than the base PCM, they were higher for n-PCM 2 during both melting and solidification. For other PCMs, experimental values were lower than the manufacturer provided values, indicating PCM aren't as capable of storing heat as they were advertised. They may be less efficient by a few percent for certain PCM and up to 44 % for others in both solidification and melting cycles.

Specific heat capacity
The specific heat capacity of PCMs determined on the basis of highest peak values are displayed in Table 46 (melting), and Table 47 (solidification). It can be seen that Cp for both melting and solidification increased slightly from the values for the base PCM for n-PCM 1 and more for n-PCM 2. It is to be noted that Cp for all PCM was given at a constant value of 2 kJ/kg·K irrespective of the PCM type in the manufacturer data sheet. From this exercise, a more appropriate value for the thermo-physical properties of the PCMs have been experimentally obtained especially for different heating rates. At this juncture, this may not seem very useful but for planning future experiments, these differences will have to be taken into consideration especially for highly sophisticated thermal regulation systems.