A Unified Total Site Heat Integration targeting method for isothermal and non-isothermal utilities

Highlights

• A new Total Site Heat Integration Targeting method is developed and presented.

• Process level targeting in the new method constrains non-isothermal utility use.

• A dairy factory, a pulp and paper mill, and a petrochemical complex are targeted.

• Comparisons suggest the conventional method over-estimates Total Site Heat Integration.

• Targets from the new method are lower but more realistic to achieve.

Abstract

This paper presents a new unified Total Site Heat Integration (TSHI) targeting methodology that calculates improved TSHI targets for sites that requires isothermal (e.g. steam) and non-isothermal (e.g. hot water) utilities. The new method sums process level utility targets to form the basis of Total Site utility targets; whereas the conventional method uses Total Site Profiles based excess process heat deficits/surpluses to set Total Site targets. Using an improved targeting algorithm, the new method requires a utility to be supplied to and returned from each process at specified temperatures, which is critical when targeting non-isothermal utilities. Such a constraint is not inherent in the conventional method. The subtle changes in procedure from the conventional method means TSHI targets are generally lower but more realistic to achieve. Three industrial case studies representing a wide variety of processing industries, are targeted using the conventional and new TSHI methods, from which key learnings are found. In summary, the over-estimation of TSHI targets for the three case studies from using the conventional method compared to new method are 69% for a New Zealand Dairy Factory, 8% for the Södra Cell Värö Kraft Pulp Mill, and 12% for Petrochemical Complex.

Introduction

Process Integration (PI) has a key role in addressing energy efficiency and waste heat improvement in process industries [1]. There are three approaches to PI: (1) graphical methods including Pinch Analysis (PA), (2) Mathematical Programming (MP) methods, and (3) hybrid approaches [2]. Application of PI techniques to a wide variety of industries has helped realise meaningful increases in energy efficiency through improved process- and Total Site-level integration [3].

PA is an elegant insight and graphical technique for Heat Integration (HI) targeting and Heat Exchanger Network (HEN) design [4]. It has been well-utilised in the process industry as a tool to maximise energy saving and heat recovery within the individual process units [5]. An important strength of the PA approach to PI is the targeting stage where important performance targets are determined prior to the design stage. Establishment of meaningful and achievable targets provides critical guidance in the design stage to the engineer of the performance limitations and inherent compromises within a system. On the other hand, Mathematical Programing typically solves network superstructures to find feasible and optimal designs. MP combines algorithmic methods with fundamental design concepts [6]. It is capable of optimising both single- and multi-objective problems including HEN retrofit [7].

Total Site Heat Integration (TSHI) was initially introduced by Dhole and Linnhoff [8] to investigate HI across plants. TSHI is a strategy for the integration of large multi-process sites to improve site-wide energy efficiency that has focused on exploiting the steam utility system to recover and place heat [9]. The method prioritises integration of individual processes and zones (i.e. defined areas of integrity [10]), before integrating across an entire site using the utility system [11]. Total Site (TS) source and sink profiles are composites of shifted Grand Composite Curves (GCC) from individual processes and applied to calculate TS targets for heat recovery, utility use, and shaft work/power generation [12]. Shortly after its initial development, Klemeš et al. [9] summarised successful applications of TSHI to an acrylic polymer manufacturing plant, several oil refineries and a tissue paper mill, which all showed utility savings between 20 and 30%. The PhD thesis of Raissi [13] presents much of the early developments of TSHI.

Inter-process integration through TSHI has recently led to increasing utility savings in slaughter and meat processing by 35% [14], large industrial parks in Japan about 53% [15] and Thailand by 28% [16], chemical processing clusters by 42% [17], and Kraft pulp mills about 13% [18]. Notable developments to the TSHI method include: temperature shifting using process [19] and stream [20] specific minimum approach temperatures, application to Locally Integrated Energy Sectors [21], integration of renewables [22], variable energy supply and demand system [23], heat exchange restrictions [24], seasonal energy availability [25], centralised utility system planning [26], retrofit investigations in TS [27], process modifications [28], minimisation of thermal oil flowrate in hot oil loops [29], heat transfer enhancement in site level heat recovery systems [30], variable energy availability [31], and TS utility system targeting [32]. There are also MP approaches to TSHI [33] including its retrofit [34].

Effectively applying TSHI techniques to processing applications and sites that required non-isothermal utilities is complex and economically challenging. The use of a Heat Recovery Loop (HRL) for site-wide heat integration of low temperature processes was investigated by Atkins et al. [35] and later formalised into a comprehensive method by Walmsley et al. [36]. In their work detailed targeting and design considerations for HRLs have included: thermal storage management [37], storage temperature selection [38], storage capacity [35], heat exchanger area sizing method [39] and performance [40], the integration of industrial solar [41], and the effect of using a nanofluid as the intermediate fluid on heat recovery [42]. Recently, Chang et al. presented the use of MINLP model with economic objectives to optimise HRLs [43].

Liew et al. [44] introduced an algebraic TS energy targeting methodology using cascade analysis. This methodology is developed mainly for targeting isothermal utilities. Non-isothermal utility (e.g. hot water and cooling water) were considered in the grassroots design [44], retrofit [27], and with variable supply and demand [31] case studies by considering only a utility supply temperature and no fixed utility target temperature. This assumption at times generates misleading energy targets for non-isothermal utility because there is no guarantee that the consumption and generation of the non-isothermal utility will maximise heat recovery. Methodology improvement, such as a targeting algorithm for non-isothermal utility, is therefore needed to systematically consider non-isothermal utility with fixed or soft target temperature.

TSHI methodologies are baseline feasibility studies for maximum energy recovery via heat exchanger and utility network. However, there are no network constraints (besides thermodynamics) in conventional TSHI for heat exchanger matches between process and utility streams. This problem was recently recognised by Sun et al. [32]. As the boiler feed water (non-isothermal utility) pinched against the TSP, they recognised that the heat may need to be transferred from multiple processes and that the network, although thermodynamically feasible, might be too complex in practice [32]. Conventional TSHI allows process-utility heat exchanger matches in series from any process for the utility to reach its target temperature and then returned to the central utility system. As a result, conventional TSHI targets for non-isothermal utilities can be overly optimistic. The HRL method, on the other hand predefines the network structure by having all HRL heat exchangers in parallel to one another (even within a process). Such a tight constraint for the network lowers the inter-plant heat recovery target and often overlooks opportunities to increase energy efficiency [45]. The gap in the literature is, therefore, the lack of an appropriate TSHI method that realistically targets both isothermal and non-isothermal utility consumption.

The aim of this paper is to introduce an improved TSHI method that calculates more realistic and achievable targets for non-isothermal utilities. The new method is referred as the Unified Total Site Targeting Method (UTST). The non-isothermal utilities in this methodology include hot water system for low temperature processes (e.g. food and beverage processing) and intermediate temperature processes (e.g. pulp and paper processing), as well as hot oil system in high temperature processes (e.g. oil refineries). The new method modified the Conventional TSHI Targeting methods, which produces substantially different TS targets for cases when non-isothermal utilities are used in the TS system. Targeting results from the new unified method are compared to the conventional TSHI method using three case studies. The effects of assumptions inherent in both methods that generate targets for TS integration are illustrated and discussed.