Performance and Reliability Improvement of Partially Shaded PV Arrays by One-time Electrical Reconfiguration

Partial shading is the most unexpected scenario encountered by the arrays that degrade the performance causing power reduction, non-convex characteristics curves, losses, hotspot, module damage, and system failure. The adoption of various reconfiguration techniques has recently provided relief to the PV array to reduce power losses during partial shading. However, these techniques exhibit vulnerabilities such as reliable operation, ease of implementation and, higher cost and complexity due to the requirement of abundant manpower, labour, complex algorithms, dynamic operation, switches and sensors that cause additional power losses. Hence, this paper presents a low cost and less complex reconfiguration technique for PV arrays to effectively increase the power generation during partial shading scenarios. The proposed reconfiguration do not require any manpower, labour, algorithms or additional devices to reduce the losses in arrays during shading. The efficacy of the technique is tested using two array sizes under various shading scenarios using MATLAB modelling and real-time field experiments. Also, for better analysis, the performance of the proposed technique is compared with conventional configurations and Sudoku reconfiguration. The investigation proclaimed that the proposed reconfiguration technique has an average power enhancement of 20% higher than any other conventional configurations.


I. INTRODUCTION
Solar Photovoltaic (PV) is one of the fastest-growing and adopted energy generation sources that utilize the energy of the sun to generate electrical energy [1]. Besides being a renewable energy source, solar PV system exhibits various other supremacy of being the reliable and noise-less generating source. However, the system encounters the major demerit of partial shading that causes severe losses and complexities while generating the maximum power equivalent to the receiving irradiance and operating temperature [2].
Partial shading mainly occurs due to various factors that include shadows of nearby infrastructures such as buildings and chimneys, trees, clouds or any foreign materials that act as a barrier between the PV module and irradiance such as leaves, dust, snow, etc. [3]. The occurrence of shading among the PV arrays reduces the performance of the arrays by minimizing the power output equivalent to the power generation of the lowest-performing or shaded modules [4]. The long time existence of partial shading in modules may result in the creation of hotspot formed by the local heat among the cells of the module [5]. This hotspot scenario can physically damage the module by detaching or melting the connectivity wires of cells, breaking the encapsulating glass, burning of cells, etc. [6]. The presence of hotspot in the module is mainly detected through various techniques such as infrared thermography [7] whose diagnosis is done by Naive Bayes classifier [8].
Generally, the modules are integrated with anti-parallel diodes to bypass the current generated by the healthy modules to prevent hotspot formation in shaded cells and enhance the power generation [9]. However, to increase the current generation of the module, turning on the bypass diodes during shading comprises the voltage of the respective modules by forcing them to operate under open-circuit mode [10]. This scenario leads to a voltage imbalance between the modules connected to the array and creates an additional complexity by forming multiple peaks in the characteristics curves of the PV array [11]. The system encounters additional losses caused by the false maximum peak tracking of the maximum power point tracking (MPPT) algorithms. To reduce the losses caused by the false maxima tracking from the P-V or power curves, various hybrid MPPT algorithms have been proposed in a wide range of literature [12]. Some examples of such techniques used for designing hybrid MPPT algorithms include Cuckoo Search [13], Particle Swarm [14], Harris Hawk [15], Grass Hopper [16], Salp Swarm [17], Hybrid Evolutionary [18], Grey Wolf [19], etc. These algorithms use various optimization techniques for searching the actual or global MPP from the non-convex characteristics curves resulting in higher power output and zero losses due to the tracking failure. However, these techniques encounter various limitations in terms of complex algorithms, powerful micro-controllers requirement, need for switches and sensors and reliability of efficient working during all shading patterns.
Arrays are formed by connecting modules in different configurations among which series configuration or long string is highly susceptible to power losses during partial shading and forms a higher number of peaks in the characteristics curves [20]. The partial shading has a puny effect on the performance of parallel-connected modules but, implementation of this configuration is practically avoided due to higher current rating that can result in losses. Hence, series-parallel (SP) configuration is widely accepted in PV power plants or roof-top systems to effective supply reliable power with desired voltage and current rating to the load. Besides SP, various other configurations such as Bridge-Linked (BL), Honeycomb (HC) and Total Cross Tied (TCT) are formed by connecting wires across the junction of the modules that act as an additional pathway for the higher current to flow without activating the bypass diodes of the shaded modules. The aforementioned configurations are widely tested in simulation and validated experimentally using different shading scenarios for reliability analysis and the study concludes that the TCT configuration has the highest potential of reducing the power losses and enhancing the generation of the array during partial shading [21].
Recently, PV array reconfiguration is gaining wide acceptance in terms of efficient shade mitigation that adopts the concept of distributing the effect of partial shading in the array and hence, reducing the mismatch losses in the system [22. The PV array reconfiguration is divided into two categories: (a) static reconfiguration, and (b) dynamic reconfiguration. In static reconfiguration, the physical position of the modules is changed to reduce the mismatch among modules whereas, in dynamic reconfiguration, the connections of the modules are changed concerning the shading pattern by using switches [23]. Some of the static reconfiguration strategies include: shade dispersion scheme (SDS) [24], fixed electrical reconfiguration (FER) [25], dominance square (DS) [26], zigzag [27], shade dispersion physical array relocation [28], shade dispersion positioning [29], Sudoku [30], improved Sudoku [31], modified Sudoku [32], competence square [33], magic square [34], etc. These strategies utilize rearrangement concepts or algorithms to change the position of the module to disperse the shading however, changing the module's position require huge manpower mainly in the case of large power plants that comprise a large number of big size modules. Similarly, various dynamic reconfiguration strategies include: genetic algorithm [35], population-based algorithm [36], two-step GA [37], two-phase array reconfiguration [38], particle swarm optimization [39], modified Harris Hawks optimization [40], swarm reinforcement learning [41], coyote optimization [42], meta-heuristic grasshopper optimization [43], etc. Also, various differential power processing based solutions for reducing the mismatch loss from the partially shaded array has been proposed in a wide range of literature [44]. Some examples include optimized PV-to-bus DPP system [45], power balancing point-based optimization algorithm [46], etc. The afore-mentioned strategies are proved as effective in mitigating the power losses due to partial shading in terms of higher power generation and reduced peak counts as compared to the conventional configurations. However, these techniques encounter the major demerit in terms of implementation as they require a higher number of switches, sensors and powerful microcontrollers that add cost, complexities and additional switching losses to the PV array. Also, these techniques are vulnerable to short-circuit fault that can occur due to the failure of any switches or controllers operating them.
So, in contrast, a reconfiguration technique for the PV arrays has been proposed in this paper that utilizes a onetime electrical connection to reduce the losses that occur due to the presence of partial shading in the system. The superiority of the proposed technique lies in some major aspects such as no sensors, switches, labour or manpower requirement for implementation. The proposed electrical connection can be implemented in PV arrays through an algorithm with the added advantage of being static. The proposed strategy is initially implemented to a 3x3 PV array and compared with the conventional configurations such as series-parallel (SP), bridge-linked (BL), honeycomb (HC), and total cross tied (TCT). Later on, a 9x9 PV array with the conventional configurations, Sudoku repositioning strategy and electrical reconfiguration is considered for comparison. The performance investigation is carried out in MATLAB software and using an experimental prototype under various shading scenarios and the comparison is done using power generation, losses, power curves analysis, efficiencies and performance ratio.

II. EXISTING SYSTEMS: MODELLING, DESCRIPTION AND PERFORMANCE PARAMETERS
In this section, the mathematical modelling of the PV array and its configurations and formulations used in the work has been briefly discussed.

A. SIMULATION MODELLING AND EXPERIMENTAL SETUP OF PV ARRAYS
The PV arrays are mainly formed by connecting numerous modules in series and parallel connections to achieve the desired voltage and current rating. The single-diode model is used to mathematically design the modules which are electrically represented by a circuit containing a current source, parallel diode, series resistance (RS) and shunt resistance (RSh) in addition to an anti-parallel connected bypass diode. The mathematical equation involved in the modelling of the PV module is represented in equation (1) where IM, VM, IPh, IO, F and VT denotes the module's maximum current, maximum voltage, photo-generated current, diode current, ideality factor and thermal voltage respectively.
The mathematically modelled modules are connected in series and parallel connections to obtain the array whose electrical circuit representation has been depicted in FIGURE 1 where 'M' and 'N' represents the number of rows and columns of the array.
The maximum power of the PV array (PA) can be determined using the equations (2) where i, j, VM, IM, VA, IA, M, and N are the rows count, columns count, individual module maximum voltage, individual module maximum current, array maximum voltage, array maximum current, number of rows and number of columns at a given irradiance and operating temperature respectively.

*BD-Bypass Diode
The specification of the PV module at standard testing condition (STC) i.e. 1000W/m 2 receiving irradiance and 25 o C operating temperature that used in the work has been given in Table I. However, to validate the experimental results, the simulation has been conducted based on the real-time environment where the PV module has received the maximum irradiance of 800W/m 2 with a module temperature of 45 o C during the healthy scenario. The specification of the PV module based on field condition parameters has been given in Table I. The prototype setup for conduction of experiments in field scenario has been illustrated in FIGURE 2 where nine modules with the rating given in Table I are used to obtain various configurations for the 3x3 PV array. The shading scenarios are acquired by applying thin plastic sheets of different colours that act as a barrier between the irradiance and receiving module surface. The PV array has been connected to a variable load (rheostat of 220Ω, and 10A) through an ammeter for current measurement and multimeter for voltage measurement. The irradiance received by the modules is measured by two solar power meters whereas an infrared thermometer is used for PV module temperature measurement. The experiment is conducted at the roof of Solar Research Lab, ITER of SOADU from 10:30 AM to 12Noon where the site received an average irradiance of 800-850W/m 2 with an ambient temperature of 34.5 o C.
In the simulation, the module has generated power of 35.26W during 800W/m 2 whereas, in the experiment, the power has been recorded as 34.96W with simulation to the experimental error of 0.85%. The simulation and experimental power generation for 600W/m 2 have been recorded as 26.10W and 25.96W respectively with an error of 0.53% whereas, for 450W/m 2 , the recorded power values for simulation and experiments are 19.33W and 19.16W respectively with an error of 0.87%. Similarly, during 200W/m 2 , the module has generated 8.12W in simulation and 7.92W in the experiment with an error of 2.42% whereas, for 100W/m 2 , the simulation and experimental powers are recorded as 3.83W and 3.75W respectively with an error of 2.08%. The study encountered a minor deviation (<5%) between the power generation data of simulation and experiments due to certain unavoidable factors such as fluctuating irradiance, temperature difference, internal cells mismatch, scratched module glass, wire losses.

1) CONVENTIONAL ARRAY CONFIGURATIONS
The most common and widely used configuration of PV array is series-parallel (SP) where the modules are initially connected in series to increase the voltage forming string and similar strings are connected in parallel to increase the current. However, the series-parallel (SP) configuration encounter a limitation of effective current dispersion during partial shading scenarios that result in severe power losses in the array.
Hence, as a solution to this limitation, the junctions of the SP configured modules are connected with wire ties to provide additional paths for the higher current to flow through preventing bypass diodes activation in the modules under shading. The Bridge-Linked (BL) and Honeycomb (HC) are the configurations with the lowest redundancy where the ties are connected across the module's junction in a bridging manner, however; in the case of total-cross-tied (TCT), the terminals of each module are connected using wire ties. The series-parallel (SP), bridge-linked (BL), honeycomb (HC) and total cross tied (TCT) configurations for 3x3 and 9x9 PV arrays have been delineated in FIGURE 3 (a), (b), (c) and (d) respectively.

1) SUDOKU RECONFIGURATION TECHNIQUE
The Sudoku reconfiguration [26] is obtained from the logicbased number-placement puzzle where the 9x9 array is divided into nine 3x3 sub-arrays with the rows and columns containing non-repeating digits from 1 to 9. The initial 9x9 array has been given in FIGURE 4 (a) whereas the final renumbered array based on the Sudoku puzzle is depicted in FIGURE 4 (b). In this technique, the physical position of the modules with TCT connection is changed based on the renumbered Sudoku pattern without altering the electrical connection. In the Sudoku technique, the modules located at a single row of TCT configuration are shifted to different rows to minimize the effect of shading occurring at a single row and dispersing it to other rows of the array that increases the current entering the node and reduces module bypassing. Taking an example, the position of module 12 (first-row second column) is changed to the seventh-row second column with the module connection remaining at the first row. Similarly, module number 33 (third-row third column) is shifted to the ninth-row third column with an electrical connection remaining at the third row. In this way, the physical position of all the modules is shifted according to the Sudoku renumbered pattern to disperse the shade throughout the array for power enhancement during shading. As the electrical connection of the array remained unchanged, so, the voltage and current rating remains equal to that of SP, BL, HC and TCT configurations.

C. Performance Parameters
The performance of the array configurations is investigated using various measuring parameters to study the efficacy during shading. The array power generation (PA) is one of the major parameters that can be used for the performance evaluation of the configurations and calculated as The percentage of power loss (PL) encountered by an array during a particular shading scenario is derived from equation (5) where PSTC indicates the array power generation at STC.
The efficiency of array configurations (ηPG) is calculated using equation (6) in which 'G' and 'A' referred to the irradiance are receiving area of the module respectively.
The power conversion efficiency (ηPC) is determined equation (7) where 'PC' is the calculated power extracted from the sum of power generation of individual PV modules.
The performance ratio (PR) of the array configurations can be as the percentile ratio of power generation during a shading scenario to unshaded case and given as

III. ONE-TIME ELECTRICAL RECONFIGURATION: DESCRIPTION, IMPLEMENTATION AND EVALUATION
The proposed one-time electrical reconfiguration is a permanent solution for partial shading prone arrays which is based on electrical reconnection of the PV modules to disseminate the intensity of partial shading in the array. The main objective of distributing the shading throughout the array is to reduce the possibility of losses caused by the lower current generation in rows of the array.
The reconfiguration initially begins with running the pseudo-code programmed using C++ language as given above that takes the number of rows (M) and columns (N) as input. The program initializes with constructing a matrix Aij of size M×N with i and j as row and column indices respectively. The output of the program will be a new remembered matrix Bij of the same size M×N where the numbering of the initially considered matrix is changed by swapping the elements across the rows and columns. After that, a TCT connected array of size M×N is taken that will act as the guiding reference to carry out the electrical connection of the electrically isolated renumbered array. The detailed steps involved in the implementation of the proposed reconfiguration to a 3x3 array are given in pictorial (FIGURE 5) and explained below. Program Execution: Step 1: Enter the number of rows and columns of the array as 3 and 3 respectively.
Step 3: After execution, the program will create a renumbered array (or matrix) based on the shade dispersion logic as shown in FIGURE 5 (c). The modules of the initial 3x3 array are considered to be electrically connected in TCT configuration (FIGURE 5 (d)) whereas the modules of the renumbered PV array are electrically isolated or unconnected with each other.

Electrical Connection:
Step 4: At first, the modules number 11, 21 and 31 are connected in series [according to the TCT configuration shown in FIGURE 5 (d)] represented by red wires for the positive terminal of module 11, green wires as a connection between module 11 negative and module 21 positive terminals, and blue wires as the connection between module 21 negative and module 31 positive terminals shown in FIGURE 5 (e).
Step 5: Secondly, modules number 12 is connected in parallel with module 11 represented by red (positive) and green (negative ) wires and then connected in series with module 22 (green wires as negative of 11 and positive of 22). Similarly, module 22 will be connected in parallel with 21 (green wires) and series with 32 (blue wires) as shown in FIGURE 5 (f).
Step 6: Thirdly, modules number 13 and 12 are connected in parallel (red wires for positive terminals and green wires for negative terminals) and then module 13 is connected in series with 23 (green wires). Similarly, PV module 23 is connected in parallel with module 22 (green wire for positive and blue wire for negative terminals) and series with 33 and then modules 33 and 32 are connected in parallel (blue wire for positive and black wire for negative terminals as shown in FIGURE 5 (g).
For connection simplicity, another approach can be adopted to improve the ease of wiring where the terminals of the PV modules are connected to specific knots of different colours that either connect modules in series or parallel configuration. Considering the connection of the 3x3 renumbered PV array as an example, the positive terminals of modules number 11, 12 and 13 are directly connected to the red knot which is the positive output terminal of the array. The negative terminals of modules 11, 12 and 13, and positive terminals of modules 21, 22 and 12 are connected to the green knot. The blue knot has the connections of positive terminals of modules 31, 32, and 33, and negative terminals of modules 21, 22 and 23. The black knot contains the negative terminals of modules 31, 32 and 33 that is connected to the negative output terminal of the PV array. The connection diagram of the renumbered modules for one-time reconfiguration and modules connected to different knots has been shown in FIGURE 6.
Since the wiring of the reconfigured PV array has been done equivalent to the connection of the TCT array, hence, the voltage, current and power ratings of the array with electrical reconfiguration remain the same. In this reconfiguration concept, the modules present in a single row and column of the TCT array are electrically reconnected in such a way that the modules will appear to be at different rows and columns which reduces the chances of shading occurring at a single row and distributes it all over the array.
The proposed technique aims to increase the current entering to the nodes during shading resulting in an improved power generation of the array.
The proposed reconfiguration strategy has been validated mathematically using a partially shaded 5x5 array and compared with the TCT array in terms of theoretical power generation and mismatch losses. The shading scenario applied to the 6x6 array with TCT configuration has been depicted in FIGURE.7 (a) where all the modules of the first row received 100W/m 2 , the second row received 400W/m 2 and other rows received 1000W/m 2 (unshaded).
The current generated by an individual module at any irradiance level can be mathematically estimated as In TCT, the modules present in a single row are connected in parallel whereas similar rows are connected in series.  Hence, the current generated by an individual row can be estimated as the sum of individual module current as given in equation (10) whereas the row voltage remains equal to the voltage of a single module.
Hence, the current generated by the first row (IR1) of the TCT array can be theoretically calculated as Similarly, the current generated by the second row (IR2) of the TCT array can be calculated as The current generated by the unshaded rows i.e. third (IR3), fourth (IR4), fifth (IR5) and sixth (IR6) of TCT configuration are estimated as Considering the TCT PV array with modules having no bypass diodes (as shown in FIGURE 7 (b)), the maximum current output (IATCT) of the array will be the current output of lowest-performing modules (series current limitation) i.e. 0.6IM whereas the maximum voltage output (VATCT) will be equal to 6VM. Hence, the maximum power output of the TCT array without bypass diodes can be theoretically calculated as In TCT array with bypass diodes (as shown in FIGURE 7 (c)), the maximum current output (IATCT) of the array will be 6IM as the current generated by the unshaded modules flows through the bypass diodes without flowing through the shaded modules. However, turning on of bypass diodes result in voltage reduction of the respective module's row (opencircuit of modules) equivalent to the forward-biased voltage of the bypass diode i.e. VD. Hence, two rows of the TCT array are forwarded generating a voltage of VD by each row resulting in a total array voltage equal to 4VM+2VD. Hence, the total power output of the TCT array with bypassing the shaded rows considering VD<<<VM and can be theoretically calculated as The shade dispersion in the case of the proposed electrical reconfiguration has been shown in FIGURE 8 (a) where the shading present over the first two rows of the TCT is distributed throughout the PV array. The maximum current generation of the rows can be mathematically calculated as Hence, the maximum power output of the PV array with the proposed one-time electrical reconfiguration (PAE) and no bypass diodes [FIGURE 8 (b)] can be calculated as Similarly, the power output of one-time reconfiguration array with bypass diodes has been calculated as But, practically turning on of the bypass diodes scenario is limited as shaded modules bypassing mainly occur when the power generated by the modules under no shade is higher than the net power generation of the array. Hence, for reduced complexity, the power generation of the PV arrays without bypassing the shaded modules is considered for comparison.
The maximum power generation of the 6x6 PV array during no shading/ healthy scenario can be calculated as The total power losses encountered by the PV array with TCT configuration (PLossT)) can be calculated as Similarly, the power losses encountered by the array with proposed one-time electrical reconfiguration (PLossE) can be calculated as Hence, from equations (21) and (22), it can be observed that the PV array with the proposed one-time electrical reconfiguration has reduced the power losses up to 21.6VMIM as compared to the TCT array. Also, it can be stated that the proposed one-time electrical reconfiguration has generated 85.71% higher power as compared to TCT configuration during this particular shading scenario which reflects the efficacy of the proposed strategy in terms of higher power generation.

IV. SIMULATION AND EXPERIMENTAL RESULTS: ANALYSIS AND DISCUSSIONS
The proposed one-time electrical reconfiguration has been tested using two different array sizes i.e. 3x3 and 9x9 and compared with various configurations such as series-parallel (SP), bridge-linked (BL), honeycomb (HC), total cross tied (TCT) and Sudoku reconfiguration (for 9x9 PV array). The entire investigation is conducted in the MATLAB software and prototype field experimental setup (for 3x3 PV array). The rating of the PV module considered for the simulation and experimental analysis of the 3x3 array has been tabulated in Table I. However, the 9x9 PV array has been studied only in the simulation environment using large size PV modules generating a maximum power output of 325W, maximum voltage: 37.8V, maximum current: 8.60A, open-circuit voltage:46.60V and short-circuit current: 9.20A at STC.
The comparison of the configurations has been done using various parameters given in equations (4)-(9) that includes power generation, mismatch loss, power loss, power generation efficiency, power conversion efficiency and performance ratio. The analysis has been carried out using different shading scenarios that differ in terms of size, strength and pattern where the modules under no shading condition are operated at 800W/m 2 irradiance and 50 o C module temperature (based on real-time environment data) and the shaded modules received different levels of irradiances i.e. 100W/m 2 , 200W/m 2 , 400W/m 2 and 600W/m 2 . The maximum power generated by the unshaded 3x3 and 9x9 arrays in the simulation at 800W/m 2 and 50 o C has been found as 317.61W and 22894.55W respectively.

A. SCENARIO A
The shading for the PV arrays with SP, BL, HC, TCT and proposed one-time reconfiguration has been depicted in FIGURE 9 (a). During this particular shading, the modules present in the first-row first-column, first-row secondcolumn, second-row first-column and second-row second column are subjected to receive lower irradiance levels of 100W/m 2 , 100W/m 2 , 200W/m 2 and 200W/m 2 respectively. The total available power from the array during this shading scenario has been calculated as 200.2W (simulation) and 198.14W (experiment). The rows currents estimation for the TCT configuration has been shown in FIGURE 9 (b) in which the theoretical currents generation of the array have been estimated as 1IM (first row), 1.2IM (second row) and 2.4IM (third row) respectively. Similarly, the theoretical rows current generated by the array with proposed one-time reconfiguration (FIGURE 9 (c)) has been calculated as 1.7IM for the first row, 1.8IM for the second row and 1.1IM for the third row.
The theoretical estimation of the array current, voltage and power based on the bypassing of shaded rows in case of TCT and proposed one-time electrical reconfiguration have been tabulated in Table II. From the table, it can be seen that the location of the global maximum power point (GMPP) of TCT is at 3VMIM.   and 158.76W respectively. Hence, from the above simulation and experimental analysis, it has been observed that the proposed one-time reconfiguration improved the array performance with a higher power output of 19.36%, 13.64%, 14.59% and 10.96% higher power than SP, BL, HC and TCT respectively.

B. SCENARIO B
The shading scenario B has been shown in FIGURE 12 (a) where all the three rows of the arrays are subjected to partial shading with the first two modules of the first-row, secondrow and third-row (from left side view) receiving lower irradiances (400W/m 2 , 600W/m 2 and 200W/m 2 respectively) whereas the unshaded modules (last modules of each row) received 800W/m 2 . The shading structure in the case of the TCT array has been shown in FIGURE 12 (b) and the shade dispersion by the one-time electrical reconfiguration has been displayed in FIGURE 12 (c) along with the rows current estimation. The theoretical currents of row 1, row 2 and row 3 in the TCT array have been estimated as 1.6IM, 2IM and 1.2IM respectively whereas the reconfigured array has rows current estimated as 1.4IM (row 1), 1.6IM (row 2) and 1.8IM (row 3). The theoretical current, voltage and power estimations of the array with TCT configuration and one-time electrical reconfiguration during scenario B have been calculated and given in Table III. It can be observed from the table that the GMPP in the case of TCT i.e. 3.2VMIM can be achieved only after bypassing one complete row of the array however, the GMPP of the array with reconfiguration has been extracted without bypassing any row of the array at 5.4VMIM. The power curves (simulation) of the PV array configurations including SP, BL, HC, TCT and proposed reconfiguration have been shown in FIGURE 13 where it can be visualized that the curve of the proposed reconfiguration (black curve) lies at the higher position indicating higher power generation than others.   The total available power from the PV array has been calculated as the sum of individual module power generation and found as 212.22W. The PV array with the electrical reconfiguration has generated a notable higher power output (196.66W) followed by the TCT (174.37W), BL (173.59W), HC (167.53W) and SP (163.89W). The SP experienced the highest losses of 153.72W (mismatch) and 63.58% (power) followed by HC (50.08W, 62.77%), BL (144.02W, 61.42%) and TCT (143.24W, 61.25%) whereas the electrical reconfiguration array encountered a minimal mismatch and power losses of 120.95W and 56.2% respectively. Taking the power generation and conversion efficiencies into account, the proposed electrical reconfiguration has the higher efficiencies (9.22% generation and 92.66% conversion) as compared to TCT (8.18% generation and 82.16% conversion), BL (8.14% generation and 81.79% conversion), HC (7.86% generation and 78.94% conversion) and SP (7.69% generation and 77.22% conversion) configurations. Hence, it can be concluded that the proposed strategy has better performance (61.91) during shading scenario B as compared to the TCT (54.90%), BL (54.65%), HC (52.74%) and SP (51.60%) configurations. Also, there is a power enhancement in the PV array with proposed reconfiguration concerning SP (19.99%), BL (13.28%), HC (17.38%) and TCT (12.18%) configurations.

C. SCENARIO C
During this shading scenario (FIGURE 14 (a)), the modules present in the first row of the array are operated under no shading scenario receiving 800W/m 2 whereas the first two modules of the second row (from the left side view) received 400W/m 2 , first two modules of third-row received 200W/m 2 and the third modules of second and third rows received 600W/m 2 .
The shading scenario C occurrence in the case of the TCT array has been delineated in FIGURE 14 (b) whereas the shade dispersion by the one-time electrical reconfiguration strategy has been shown in FIGURE 14 (c).

FIGURE 15. Power curves of PV arrays (SP, BL, HC, TCT configurations) and One-time electrical reconfiguration) during partial shading scenario C (simulation analysis)
The row's current estimation of the TCT array has been mathematically calculated as 2.4IM for the first row, 1.4IM for the second row and 1IM for the third row. Similarly, the row current estimation of the array with proposed reconfiguration for shading scenario C has been mathematically calculated as 1.6IM, 1.4IM, 1.8IM for the first, second and third rows respectively. The theoretical current, voltage and power calculation of the array with TCT configuration and proposed reconfiguration has been tabulated in Table IV. From the table, it can be observed that the PV array with TCT configuration has the theoretical GMPP at 3VMIM whereas the proposed reconfiguration has a notable higher GMPP lying at 5.4VMIM.
The power curves of the PV array are represented in (TCT). Also, power generation and conversion of efficiencies of the TCT (6.81%and 68.46%) is higher as compared to SP (6.53% and 65.66%), BL (6.78% and 68.08%) and HC (6.59% and 66.20%). However, the proposed reconfiguration strategy encountered the lowest mismatch (122.95W) and power (56.74%) losses with higher power generation (9.13%) and conversion (91.72%) efficiencies. Hence, it can be stated that the proposed strategy has higher performance (61.28%) as compared to the TCT (45.74%), BL (45.49%), HC (44.23%) and SP (43.87%) configurations. The proposed strategy has significantly generated 39.69% higher power than SP, 34.71% than BL, 35.50% than HC and 32.06% than TCT during this particular shading scenario.

D. SCENARIO D
The shading scenario D is shown in FIGURE 16 (a) in which the whole array is under shading receiving lower irradiances of 100W/m 2 , 200W/m 2 , 400W/m 2 and 600W/m 2 . This scenario is considered to be one of the complex scenarios as besides the available irradiance of 800W/m 2 , the PV array operates under more than three irradiance levels. The rows currents in TCT configuration have been theoretically calculated as 1IM (Row 1), 0.4IM (Row 2) and 1.8IM (Row 3) as shown in FIGURE 16 (b). Similarly, theoretical currents generated by the array after shade dispersion by the proposed strategy for row1, row 2 and row 3 have been calculated as 1.6IM, 1.4IM and 1.8IM. The theoretical estimation of row currents, voltage and power of the PV arrays with TCT configuration and proposed reconfiguration are calculated and tabulated in Table V. It can be observed that the GMPP (2VMIM) of TCT can be acquired after bypassing one row of the array whereas the GMPP of the array with the proposed strategy occurred without bypassing any row i.e. 4.2VMIM.    respectively. The power generation of the array with the proposed strategy is 6.47% higher than SP, 6.60% than BL, 4.70% than the HC and 4.41% than TCT configurations.

F. SCENARIO F
During scenario F (FIGURE 19 (a)), the PV modules present over the first and second rows of the array have been subjected to shading receiving lower irradiance levels of 100W/m 2 whereas the modules of the third row received 800W/m 2 (no shading). The total available power in the array under this shading is found as 128.76W. The power curves of the arrays have been shown in FIGURE 19 (b) representing that conventional configurations have equal power output i.e. 94.82W with mismatch loss as 222.79W, power loss as 78.92%, power generation efficiency as 4.44%, power conversion efficiency as 74% and performance ratio as 29.85%. However, the power output of one-time electrical reconfiguration is found to be higher i.e. 127.13W with lower losses of 189.48W (mismatch) and 71.52% (power). Also, there is a higher power output and conversion efficiencies of 6.01% and 98.75% respectively with a performance ratio of 40.02%. The PV array with proposed reconfiguration has 34.07% power enhancement than the SP, BL, HC and TCT configurations.
The parameters comparison of the proposed technique with the SP, BL, HC and TCT configurations during all the above-studied scenarios has been graphically represented in FIGURE 20. The graphs clearly state that the proposed reconfiguration excelled in performance in terms of reduced losses and improved efficiencies during partial shading.
The comparative data obtained from the simulation and experiments for different 3x3 PV array configurations during shading scenarios A to F have been tabulated in Table VI. The results indicate that the array implemented with the proposed one-time electrical reconfiguration has shown excellent performance during all the shading scenarios.
Also, for enhanced analysis, several other partial shading scenarios for the 3x3 PV array have been considered. The shading scenarios, corresponding P-V characteristics curves and power generation of the conventional configurations and proposed electrical reconfiguration have been given in Table  VII. The shading scenarios considered in the table replicates the possible occurrence in a real-time scenario due to the presence of the trees and buildings shadows. From the results given in Table VII, it can be observed that the electrical reconfiguration has significantly enhanced the power generation of the array during all the shading cases as compared to conventional configurations. Also, it has been observed that in some of the shading cases, the SP, BL, HC and TCT configurations have nearly equal power generations whereas there is a notable higher power output in the case of the proposed technique. Hence, it can be summarised from the results that the adoption of the proposed technique can significantly enhance the power generation of the 3x3 PV array during partial shading.

Shading Scenarios
Power Loss (in W)

Power Conversion Efficiency
Shading Scenarios   Similarly, to prove the efficacy of the proposed one-time electrical reconfiguration, a 9x9 array has been considered for investigation. The 9x9 array has been implemented with SP, BL, HC and TCT along with Sudoku reconfiguration and one-time electrical reconfiguration. The implementation of the Sudoku reconfiguration has been pictorially explained in   During no shading scenario, the 9x9 PV arrays are operated at 800W/m 2 and 50 o C generating the maximum power of 2858.57W whereas the shaded modules are operated at 100W/m 2 . The shading case A for the 9x9 PV array configurations is shown in Similarly, the mismatch and power losses of the 9x9 PV array with electrical reconfiguration have been calculated as 5279.35W and 33.08% and found to be significantly lower as compared to the other configurations. The performance summary of the 9x9 PV array during shading case A, case B and shading case have been tabularized in Table VIII. The above-conducted analysis gives clear evidence of the efficient functionality and shading mitigation ability of the one-time electrical reconfiguration in increasing the maximum power output and performance of the PV arrays. This results in a significant reduction in losses, improved efficiency and higher reliability of the array operating under shading. The major merits of adopting the proposed strategy are:

168.36W
 Higher power generation during shading  Lowers the losses from the array  Improved efficiency and performance during shading  No physical repositioning of modules and hence, no manpower is required  Ease of implementation in all types of PV systems  Zero switches and sensors requirement  No extra wire requirement (equivalent to TCT)  Low cost and user-friendly.  Ease fault diagnosis  Effective shade mitigation  Highly Reliable Hence, the proposed strategy can be easily implemented in PV arrays as a cost-effective solution for reducing the power loss due to the presence of partial shading.

V. CONCLUSION
In this work, a one-time electrical reconfiguration for partial shading prone PV arrays has been proposed that can act as a fixed solution to reduce the losses from the arrays. The proposed reconfiguration has been tested for two PV array sizes (3x3 and 9x9) under different shading scenarios in MATLAB and real-time experimental environments. The investigation concludes that the proposed reconfiguration strategy has generated significantly higher power as compared to the series-parallel, bridge-linked, honeycomb, total cross tied and Sudoku reconfiguration. During partial shading. Also, the proposed strategy can efficiently reduce the losses and increase the overall efficiencies and performance of the PV arrays during partial shading as compared to other configurations. The average power enhancement of the proposed electrical reconfiguration during partial shading as compared to the series-parallel, bridge-linked, honeycomb and total-cross-tied configurations has been found as 25.84%, 22.69%, 23.66% and 21.69% respectively. Hence, the proposed electrical reconfiguration is a highly reliable solution for effective power enhancement and loss mitigation in PV arrays. The reconfiguration is easy to execute in partial shading prone PV arrays that require no modules position replacement, switches and sensors, extra wires and hence, no additional cost and complexities in the system.