Large-scale Heat Exchanger Network Research Articles

A novel optimization framework integrating multiple initialization, automatic topologization and MINLP reduction to accelerate large-scale heat exchanger network synthesis

Ensuring the solution feasibility and achieving an acceptable solution time for the non-deterministic polynomial-time hard (NP-hard) problem of heat exchanger network (HEN) synthesis is still challenging. In this study, we developed a novel optimization framework for efficiently solving large-scale HEN synthesis problems to obtain near global optimal solutions, simultaneously addressing the solution's infeasibility and time. The optimization framework consists of a multiple initial HEN design model (M-INI model), a new distance-based auto structure transformer (A-STR transformer), and a reduced non-linear programming model (R-NLP model). The M-INI model generates various HEN designs based on pinch analysis by varying the minimum approach temperature and the modes of stream splitting and mixing. The A-STR transformer is first proposed to automatically convert various HEN designs into associated superstructure-based topologies as multiple initial starting points for accelerating the solution of the R-NLP model. A strategy for approximating the temperature difference calculation is further applied to reduce the computational loads. Four case studies were conducted using the proposed method to minimize the total annual cost (TAC). The solutions of general cases 1 and 2 are almost equal to the best results in the literature, while the solution time is reduced by 98.7% and 99.8%, respectively, which demonstrates the performance of the proposed method. Especially, the method was applied to two practical industrial HEN cases: a large-scale low-grade heat recovery system and a crude distillation unit. The results of large-scale case 3 show that the optimal solution near the global optimal solution only requires 251.7 s. Case 4 includes pre-splitting in scenario A and free splitting in scenario B. Compared to scenario A, the TAC and solution time in scenario B were reduced by 33.9% and 93.1%, indicating that the pre-splitting strategy excludes the optimal from the search space. The proposed novel optimization framework demonstrates the efficiency and applicability of handling large-scale HEN synthesis problems. It can also guide engineers in the design of a large-scale HEN with lower costs and effort.

Node dynamic adaptive non-structural model for efficient synthesis of heat exchanger networks

Balancing the trade-off between structural diversity (solution space) and optimization efficiency is an urgent challenge in the modeling and global optimization of heat exchanger network synthesis problems. This paper presents a node dynamic adaptive non-structural model without stream splitting to enhance optimization efficiency under the premise of ensuring an adequate solution space. The node-based non-structural stream matching mechanism can be used to make the solution space scalable and achieve the efficacy of multi-stage stage-wise superstructure with a finite number of nodes. Two node dynamic adaptive strategies (uniformly distributed and random) are established to periodically adjust the number of nodes in process streams and the node distribution of existing heat exchanger units, thereby preventing a decrease in computational efficiency due to an excess number of preset nodes while still satisfying the solution space requirements. The node dynamic adaptive non-structural model is optimized using the random walk algorithm with compulsive evolution. The proposed method was applied to four well-known heat exchanger network case studies. High-quality solutions were obtained, and they are more economical than most of the optimal results previously reported in the literature. These results demonstrate the enhanced computational efficiency and applicability of the proposed synthesis method based on a node dynamic adaptive non-structural model for efficiently solving large-scale heat exchanger network synthesis problems.

A novel two-step synthesis method with weakening strategy for solving large-scale heat exchanger networks

Efficiently solving large-scale heat exchanger network (HEN) synthesis problems remains a challenging task mainly due to the presence of discontinuity caused by integer variables as well as nonlinearity of the objective cost function. This paper focuses on efficiently solving large-scale HEN synthesis problems based on a chessboard presentation with no splits, and aims to achieve globally near-optimal solutions by using a two-step synthesis method along with a weakening strategy for fixed investment costs that is conducive to bypass the discontinuity. The first step is to generate the most promising network structure by using an improved cuckoo search algorithm to optimize integer and continuous variables simultaneously. In the second step, a cuckoo search algorithm is utilized to directly optimize the heat exchanger area of the optimal structure obtained in the first step. Meanwhile, fixed-cost weakening strategies are applied in the above two steps for improving the optimization efficiency. The proposed method has been tested on several HEN cases, and two of which, namely the 20- and 39-stream problems, are presented in this paper. For these two cases, although the solutions obtained do not contain stream splits, their total annual costs are close to best solutions with stream splits and lower than those best solutions with no-splits reported in the literature, and their chessboard network configurations are clear and simple. Based on these results, the developed method demonstrates the applicability to handle large-scale HEN problems, and is also useful for guiding the grassroots or retrofit design of large-scale HENs or other engineering process syntheses with less cost and less pollution.

Automated retrofit targeting of heat exchanger networks

The aim of this paper is to develop a novel heat exchanger network (HEN) retrofit method based on a new automated retrofit targeting (ART) algorithm. ART uses the heat surplus-deficit table (HSDT) in combination with the Bridge Retrofit concepts to generate retrofit bridges option, from which a retrofit design may be formulated. The HSDT is a tabular tool that shows potential for improved re-integration of heat source and sink streams within a HEN. Using the HSDT, retrofit bridges—a set of modifications that links a cooler to a heater to save energy—may be identified, quantified, and compared. The novel retrofit method including the ART algorithm has been successfully implemented in Microsoft ExcelTM to enable analysis of large-scale HENs. A refinery case study with 27 streams and 46 existing heat exchangers demonstrated the retrofit method’s potential. For the case study, the ART algorithm found 68903 feasible unique retrofit opportunities with a minimum 400 kW·unit–1 threshold for heat recovery divided by the number of new units. The most promising retrofit project required 3 new heat exchanger units to achieve a heat savings of 4.24 MW with a favorable annualised profit and a reasonable payback period.

Two-step MILP/MINLP approach for the synthesis of large-scale HENs

Although different methodologies for the synthesis of heat exchanger network (HEN) problems have been introduced in the last forty years, there are still significant challenges to be addressed, such as solving large-scale problems. This study focuses on synthesizing large-scale HENs using mathematical programming to achieve near globally optimal solutions based on a two-step MILP/MINLP approach. In the first step a mixed-integer linear programming (MILP) model, TransHEN, is used that obtains a globally optimal solution at selected ΔTmin. By utilisation of this model, the most promising matches are selected based on feasibility and viability. The second step entails using the matches selected in the TransHEN of Step 1 in a mixed-integer nonlinear programming (MINLP) model, HENsyn, using a reduced superstructure, to generate a feasible HEN. This study presents also a simultaneous Total Site synthesis with direct heat transfer between processes, and is the first step in the wider project of synthesising an entire Total Site with direct and indirect heat transfer; and is the first step in the wider scope of synthesising an entire Total Site with direct and indirect heat transfer; however, in order to attain this goal, a tool capable of an appropriately handling large number of streams is required. The newly developed procedure has been tested on several case studies, two of which are presented in this paper. For Case study 1 the results obtained lie within the range of best solutions obtained by other authors. Case study 2, consisting of 173 process stream and involving multiple hot utilities, shows the applicability of the developed method to handle large-scale HEN problems.

Area-based optimization approach for refinery heat exchanger networks

This paper presents a new approach to optimize large-scale heat exchanger networks (HENs). The approach is particularly suitable for operational optimization and retrofit of industrial crude oil preheat trains, but it can be applied to HENs from other processes. Two main features distinguish this approach: first, an area-based simulation model that significantly simplifies the optimization, while capturing the details of the existing HEN. Second, a decomposition procedure to divide the large-scale HEN into two simpler HENs, which are optimized sequentially. This approach includes additional features that provide a more realistic representation of preheat trains, such as temperature-dependent heat capacities, the dependence of heat transfer coefficients on flow rate variations, and new types of stream splitters and mixers for the distillation products in the HEN. An industrial case study illustrates the application of the methodology, showing that the approach identifies opportunities to reduce energy consumption with minimal changes to the HEN structure.

Costs and environmental impacts multi-objective heat exchanger networks synthesis using a meta-heuristic approach

Heat Exchanger Network (HEN) synthesis approaches based on total annual costs (TAC) optimization can reduce CO2 emissions and electricity used in plants, since the use of utilities is reduced with a cost-optimal design. However, environmental impacts (EI) are not explicitly addressed in such methods. Life cycle assessment (LCA) metrics can quantify EI and can be used as objective function. This work proposes a meta-heuristic approach to perform a multi-objective optimization (MOO) of TAC and EI in medium and large-scale HEN. The method developed efficiently achieves near-Pareto fronts for four industrial-size case studies. Optimal-TAC and optimal-EI solutions, as well as configurations with low EI but still competitive TAC are presented. Cost-optimal solutions reported in the literature are compared with non-dominated solutions obtained in this work with similar TAC. In all cases, the approach developed is able to achieve solutions with remarkably lower EI, demonstrating the importance of multi-criteria optimization in HEN synthesis.

An efficient method based on the uniformity principle for synthesis of large-scale heat exchanger networks

The optimal design of large-scale heat exchanger networks is a difficult task due to the inherent non-linear characteristics and the combinatorial nature of heat exchangers. To solve large-scale heat exchanger network synthesis (HENS) problems, two dimensionless uniformity factors to describe the heat exchanger network (HEN) uniformity in terms of the temperature difference and the accuracy of process stream grouping are deduced. Additionally, a novel algorithm that combines deterministic and stochastic optimizations to obtain an optimal sub-network with a suitable heat load for a given group of streams is proposed, and is named the Powell particle swarm optimization (PPSO). As a result, the synthesis of large-scale heat exchanger networks is divided into two corresponding sub-parts, namely, the grouping of process streams and the optimization of sub-networks. This approach reduces the computational complexity and increases the efficiency of the proposed method. The robustness and effectiveness of the proposed method are demonstrated by solving a large-scale HENS problem involving 39 process streams, and the results obtained are better than those previously published in the literature.

Efficient Method for Flexibility Analysis of Large-Scale Nonconvex Heat Exchanger Networks

A novel method is proposed for flexibility analysis of nonconvex heat exchanger networks. In this method, the direction matrix is introduced to describe the deviation direction of the uncertain parameter. Then by searching the critical directions that restrict the process flexibility with the simulated annealing algorithm, the flexibility index can be obtained. Since the directions are no longer limited to the vertices, this method can well deal with the nonconvex problems. Moreover, for large-scale problems, a decoupling strategy is developed to enhance the efficiency. On the basis of this, the entire network is decomposed into several independent subnetworks, and the one with the lowest flexibility decides the global flexibility index. Two examples are studied, and the results well demonstrate the various aspects dealt with in this work.

A Procedure for the Retrofitting of Large-Scale Heat Exchanger Networks for Fixed and Flexible Designs applied to Existing Refinery Total Site

A Procedure for the Retrofitting of Large-Scale Heat Exchanger Networks for Fixed and Flexible Designs applied to Existing Refinery Total Site

Minimizing investment cost for multi-period heat exchanger network retrofit by matching heat transfer areas with different strategies

Multi-period heat exchanger network (HEN) retrofit is usually performed by targeting and matching heat transfer areas. In this paper, based on the reverse order matching method we proposed previously, three strategies of matching heat transfer areas are proposed to minimize the investment cost for the retrofit of HEN in multi-period, in which replacement of heat exchangers, addition of heat exchangers and addition of heat transfer areas are performed. We demonstrate the procedures through three scenarios, including maximum number of substituted heat exchangers after retrofit, minimum additional heat transfer areas in the retrofitted HEN, and minimum investment cost for retrofit. The strategies are extended to a single period HEN retrofit problem. The results of multi-period and single period HEN retrofit problems indicate the effectiveness of the strategies. Moreover, these results are better than those reported in literature. The strategies are simple and easy to implement, which are of great benefit to large-scale HEN retrofit in practice.

Synthesis of large-scale multi-stream heat exchanger networks using a stepwise optimization method

In this paper, an efficient stepwise approach is presented for synthesizing large-scale multi-stream heat exchanger networks (MSHENs). The proposed approach employs the pseudo-temperature enthalpy diagram method to gain an initial network. In the next stage, several heuristic strategies are raised to conduct the selection and mergence of enthalpy intervals. The sub-networks are re-designed in the merged intervals based on the introduced strategies to disconnect the heat exchanger loops caused by involving too many intervals in the initial networks. A corresponding mathematical model is established with the objective of minimum total annual cost (TAC), and the simulated annealing and genetic algorithm (SA/GA) is adopted for the optimization. Two cases taken from literatures are studied to illustrate the effectiveness of the proposed method on solving MSHEN synthesis problems. The results obtained are better than those of previous paper and have shown an excellent performance of the multi-stream heat exchanger.

Optimisation for the retrofit of large scale heat exchanger networks with different intensified heat transfer techniques

Intensified heat transfer (IHT) techniques have recently been used for retrofit in the process industry, leading to significant energy saving in heat exchanger network (HEN) by facilitating heat transfer intensification without network topology modification. In this paper, an optimisation method has been developed for dealing with the retrofit of large scale HENs in which the location of intensified heat transfer within the network and its degree of intensification are systematically identified, given the objective function and design constraints, including topological limitation in the existing heat recovery systems. The optimisation framework developed is based on iterative optimisation of a relatively simple mixed integer linear programming (MILP), which can effectively deal with computational difficulties associated with nonlinearity. In the retrofitted HENs, several conventional intensified heat transfer techniques are available, including tube-side intensification (twisted-tape inserts, coiled-wire inserts and internal fins), and shell-side intensification (external fins and helical baffles). Suitable exchangers can be selected for enhancement by implementing one or more intensification techniques to increase the whole network energy recovery within very low retrofit cost. A large-size industrial case study is considered to demonstrate the validity and efficiency of the proposed optimisation approach.

Solving Large-Scale Heat Exchanger Network Synthesis Problems Using Particle Swarm Optimization

Heat exchanger network synthesis has been one of the most popular subjects in process design over the last 50 years. Various studies and optimization techniques have been proposed for designing optimal network with minimum total annual cost. Simultaneous synthesis approach via mathematical programming aims to find the optimal network without decomposition, which has been paid more attentions on the research recently. However, these methods might be not solvable or inefficient for large-scale problems. This paper makes an attempt to construct simultaneous synthesis model with split streams and to develop an efficient optimization framework based on particle swarm optimization for large-scale heat exchanger network synthesis problems. One example including 20 process streams is solved to give an illustration of the method.

Synthesis of Large-Scale Heat Exchanger Network Based on Sub-Networks

The optimization of heat exchanger networks (HEN) is a typical MINLP problem. For large-scale HEN, it is difficult to solve this problem globally. After optimization, the large-scale HEN is divided into several independent sub-networks automatically. The sub-network is defined as a part of the HEN in which the streams have no heat transfer with the streams outside the sub-network. If a HEN can be divided into two or more sub-networks, then, these sub-networks are independent from each other. Based on optimization of sub-networks, a new method which can solve large-scale HEN problem efficiently is proposed.

Efficient synthesis of large-scale heat exchanger networks using monogenetic algorithm

A promising method for the efficient design of large-scale heat exchanger networks is the genetic algorithm. The heat exchanger network is divided into sub-networks which are optimized by a hybrid genetic algorithm. In additional steps these sub-networks are optimized by a monogenetic algorithm. An example consisting of 39 process streams was considered in detail. Significant improvements were made in reduction of total annual costs and the number of heat exchangers.

A monogenetic algorithm for optimal design of large-scale heat exchanger networks

The optimal design of large-scale heat exchanger networks is a quite difficult task not only due to its non-linear characteristics but also due to a great number of local optima in its solution space. An explicit analytical solution of stream temperatures for the superstructure heat exchanger networks was developed, which reduces number of decision variables significantly. Based on this solution, a mathematical model for synthesis of heat exchanger networks was formulated for searching the optimal configuration of a heat recovery system by a hybrid genetic algorithm. For large-scale heat exchanger networks, a monogenetic algorithm based on the optimization of sub-networks is proposed. In the first step of the optimization, the hybrid genetic algorithm is applied to the synthesis of the whole heat exchanger network for finding the functional groups (sub-networks) rather than the chromosomes (positions of the heat exchangers and splits of the streams) and genes (areas and heat capacity flow rates). Then the monogenetic algorithm for evolution of the functional groups is carried out to improve the HEN. This procedure was applied to examples taken from literature and better results were obtained.

Synthesis of Large-scale Multistream Heat Exchanger Networks Based on Stream Pseudo Temperature

Effective temperature level of stream, namely stream pseudo temperature, is determined by its actual temperature and heat transfer temperature difference contribution value. Heat transfer temperature difference contribution value of a stream depends on its heat transfer film coefficient, cost per unit heat transfer area, actual temperature, and so on. In the determination of the suitable heat transfer temperature difference contribution values of the stream, the total annual cost of multistream heat exchanger network (MSHEN) is regarded as an objective function, and genetic/simulated annealing algorithm (GA/SA) is adopted for optimizing the heat transfer temperature difference contribution values of the stream. The stream pseudo temperatures are subsequently obtained. On the basis of stream pseudo temperature, optimized MSHEN can be attained by the temperature-enthalpy ( T-H) diagram method. This approach is characterized with fewer decision variables and higher feasibility of solutions. The calculation efficiency of GA/SA can be remarkably enhanced by this approach and more probability is shown in searching the global optimum solution. Hence this approach is presented for solving industrial-sized MSHEN which is difficult to deal by traditional algorithm. Moreover, in the optimization of stream heat transfer temperature difference contribution values, the effects of the stream temperature, the heat transfer film coefficient, and the construction material of heat exchangers are considered, therefore this approach can be used to optimize and design heat exchanger network (HEN) with unequal heat transfer film coefficients and different of construction materials. The performance of the proposed approach has been demonstrated with three examples and the obtained solutions are compared with those available in literatures. The results show that the large-scale MSHEN synthesis problems can be solved to obtain good solutions with the modest computational effort.

06/00311 Solving large-scale retrofit heat exchanger network synthesis problems with mathematical optimization methods: Björk, K.-M. and Nordman, R. Chemical Engineering and Processing,2005, 44, (8), 869–876

06/00311 Solving large-scale retrofit heat exchanger network synthesis problems with mathematical optimization methods: Björk, K.-M. and Nordman, R. Chemical Engineering and Processing,2005, 44, (8), 869–876

Synthesis of large-scale heat exchanger networks using a sequential match reduction approach

In this paper, a sequential approach is outlined which generates networks for industrial sized heat exchanger network synthesis (HENS) problems. The proposed match reduction approach solves a sequence of subproblems, posed as transportation problems, successively reducing the set of matches that are considered in the next stage. The terms involved in the objective function are included at each step with increasing accuracy, until a final design is obtained. One of the subproblems identifies subsets of matches that can be designed separately. Due to the sequential approach, the final design is an approximating solution. Two examples are presented to illustrate the potential of the proposed method. The results show that large HENS problems can be solved to good solutions with modest computational effort. The obtained solutions are in fact, better than the results reported earlier in the literature.