Dr Michael Short

Mass and heat integration are important to achieving economically and environmentally sustainable processes through increased efficiency. Typically, heat and mass exchange networks are solved separately using process integration techniques such as pinch technology or formulating nonconvex mixed-integer nonlinear programming (MINLP) problems, which are challenging to solve. To simplify the MINLP, shortcut models are employed, which can result in under/overestimation of the real network, leading to suboptimal or infeasible designs. We introduce a new optimisation algorithm for combined heat and mass exchanger network synthesis (CHAMENS), including detailed design models. The method uses shortcut models in an MINLP to find network topology, followed by a nonlinear programming (NLP) suboptimisation. The NLP allows non-isothermal and non-isocompositional mixing, uses detailed unit models of packed columns based on orthogonal collocation on finite elements (OCFE), and detailed shell and tube heat exchanger designs. We incorporate a differential-algebraic equation (DAE) based shell and tube heat exchanger design model via surrogates in a trust region filter (TRF) framework, guaranteeing optimal solutions for the detailed exchanger models are found by the surrogate models. We demonstrate the proposed approach on a case study, showcasing its performance and the need to incorporate detailed unit models in topology optimisation to find practical optimal designs.

Optimisation models for the design of distributed energy systems (DES) often exclude inherent nonlinearities and constraints associated with alternating current (AC) power flow and the underlying distribution network. This study aims to assess this gap by comparing the performance of linear and nonlinear formulations of DES design models, connected to and trading with an AC grid. The inclusion of the optimal power flow (OPF) constraints within the DES design framework is demonstrated in the methodology. A residential case study is used to test both models and compare the designs obtained from the two formulations. The results highlight that DES designs obtained are different when constraints related to the underlying distribution network are added, particularly when electricity storage is not considered. Overall, this study highlights the need for future modelling efforts to include OPF within DES optimisation frameworks to obtain practically feasible designs, rather than considering them as standalone problems.

•Detailed heat exchanger models are embedded within a heat exchanger network synthesis optimization.•A trust region filter algorithm is used as a surrogate modelling strategy for HENS for the first time to incorporate pressure drops, numbers of shells, etc.•An integer-cut strategy algorithm is introduced to explore different topologies.•Results are compared to existing methods for solving HENS with detailed models, demonstrating excellent performance. We develop a trust region filter strategy for simultaneous optimal design of heat exchanger networks that includes detailed design of shell-and-tube heat exchangers. The strategy first solves a mixed-integer nonlinear programming (MINLP) formulation with shortcut models to generate candidate network topologies, which are then used in a non-isothermal mixing nonlinear programming (NLP) suboptimization with detailed optimal exchanger design models embedded using a modified trust region filter (TRF) algorithm. An integer cut based strategy is used to bound the solutions from MINLP and the NLP which aids in convergence to the solution of the overall simultaneous design problem. Under assumptions, the TRF based strategy can guarantee convergence to near optimal solutions of the overall design problem. The presented solution strategy is thus able to find optimal heat exchanger network designs based on the simultaneous optimization of the network topology and mass and energy balances, together with detailed shell-and-tube heat exchanger optimization, including the number of shell and tube passes, pressure drops, and tubes, tube lengths, etc. The proposed strategy is tested on three literature based case studies and their results are compared with previous studies to showcase its performance.

A new method for the detailed design of shell and tube heat exchangers is presented through the formulation of coupled differential heat equations, along with algebraic equations for design variables. Heat exchanger design components (tube passes, baffles, and shells) are used to discretize the differential equations and are solved simultaneously with the algebraic design equations. The coupled differential algebraic equation (DAE) system is suitable for numerical optimization as it replaces the nonsmooth log mean temperature difference (LMTD) term. Discrete decisions regarding the number of shells, fluid allocation, tube sizes, and number of baffles are determined by solving an LMTD‐based method iteratively. The resulting heat exchanger topology is then used to discretize the detailed DAE model, which is solved as a nonlinear programming model to obtain the detailed exchanger design by minimizing an economic objective function through varying the tube length. The DAE model also provides the stream temperature profiles inside the exchanger simultaneously with the detailed design. It is observed that the DAE model results are almost equal to the LMTD‐based design model for one‐shell heat exchangers with constant stream properties but shows significant differences when streams properties are allowed to vary with temperature or the number of shells are increased. The accuracy of the solutions and the required computational costs show that the model is well suited for solving heat exchanger network synthesis problems combined with detailed exchanger designs, which is demonstrated in Part 2 of the paper.

We propose a new strategy to synthesize heat exchanger networks with detailed designs of individual heat exchangers. The proposed strategy uses a multistep approach by first obtaining a heat exchanger network topology through solving a modified version of the mixed integer nonlinear programming (MINLP) stage‐wise superstructure of Yee and Grossmann, which includes a smoothed LMTD approximation and pressure drops. In a second nonlinear programming (NLP) suboptimization step, we allow for nonisothermal mixing to solve problems with or without exchanger bypasses. The selected heat exchangers along with the mass and energy balances obtained are then used to design the network with detailed exchanger designs through solving a sequence of NLPs for individual heat exchanger designs. The NLPs are based on the detailed discretized optimization models of Kazi et al., which solve quickly and reliably to obtain heat exchangers based on rigorous, first‐principles derived coupled differential equations. These models solve a differential algebraic equation system and do not rely on usual assumptions associated with other heuristic‐based exchanger design methods, such as log mean temperature difference and FT correction factors. These detailed exchanger designs are then used to update the network optimization model through sets of correction factors on heat exchanger area, number of shells, heat transfer coefficients, and pressure drops of each exchanger design, in a method based on that of Short et al. The method solves reliably, guaranteeing feasible exchangers for every potential network generated by the shortcut models, through validation with rigorous heat exchanger models at every iteration. In addition, the method does not increase the nonlinearity of the MINLP model, nor does it require any manual intervention or initialization from the user. Three examples are solved and the results are compared to those obtained in the literature.

Multivariate spectroscopic data is increasingly abundant in the chemical and pharmaceutical industries. However, it is often challenging to estimate reaction kinetics directly from it. Recent advances in obtaining kinetic parameter estimates from spectroscopic data based on large-scale nonlinear programming (NLP), maximum likelihood principles, and discretization on finite elements lead to increased speed and efficiency (Chen et al., 2016). These new techniques have great potential for widespread use in parameter estimation. However they are currently limited due to their applicability to relatively small problem sizes. In this work, we extend the open-source package for estimating reaction kinetics directly from spectra or concentration data, KIPET, for use with multiple experimental datasets, or multisets (Schenk et al., 2020). Through a detailed initialization scheme and by taking advantage of large-scale nonlinear programming techniques and problem structure, we are able to solve large problems obtained from multiple experiments, simultaneously. The enhanced KIPET package can solve problems wherein multiple experiments contain different reactants and kinetic models, different dataset sizes with shared or unshared individual species’ spectra, and can obtain confidence intervals quickly based on the NLP sensitivities. In addition, we propose a new variance estimation technique based on maximum likelihood derivations for unknown covariances from two sample populations. This new variance estimation technique is compared to the previously proposed iterative-heuristics-based algorithm of Chen et al. (2016) for distinguishing between variances of the noise in model variables and in the spectral measurements. We demonstrate the new techniques on a variety of example problems, with sample code, to show the utility of the approach and its ease of use. We also include the curve-fitting problem to cases where we have concentration data given directly, and are required to estimate kinetic parameters across multiple experimental datasets.

The year 2020 has seen the emergence of a global pandemic as a result of the disease COVID-19. This report reviews knowledge of the transmission of COVID-19 indoors, examines the evidence for mitigating measures, and considers the implications for wintertime with a focus on ventilation.

Mass exchanger networks (MENs) are used to remove/recover contaminants from polluted streams through absorption with available process streams or external mass separating agents. Process Integration techniques such as Pinch Technology (PT) or mathematical optimisation can be used to synthesise optimal networks, however a lack of accessible software and difficulties in formulating the non-convex problem has stunted research. This article presents an open-source Python package for the synthesis of optimal MENs. The package uses the algebraic modelling language, Pyomo, and takes advantage of Python’s object-oriented nature to solve a series of optimisation problems, improving on the performance of previous approaches to the problem of incorporating detailed unit designs into MEN synthesis. The package uses automated initialisation strategies to first solve a superstructure-based mixed-integer nonlinear program (MINLP). Thereafter, a detailed optimisation model, formulating the packed column as a system of differential-algebraic equations, is used to design the columns. This detailed packed column design is used to update the MINLP through correction factors, driving the network solution towards the detailed unit optimisation solutions. The new software, called MExNetS, implements this strategy in a user-friendly package that is easily modified and well-documented. In addition to the new software implementation, novel strategies are employed to ensure feasibility at each iteration, which is a challenge in these non-convex optimisation formulations, and new binary cuts are generated and applied to the MINLP that can significantly speed up convergence compared to the previous study. The package also contains automatic superstructure generation based on user-inputted data, with the hope that this software can inspire further research in this area and be accessible to practitioners.

This chapter presents a synthesis method for multi‐period combined heat and mass exchange networks (CHAMENs). The mass exchange network involves a regeneration network (REN) superstructure with multiple recyclable mass‐separating agents (MSAs) and multiple regenerating streams. The CHAMENs superstructure with REN is integrated with utilities generated from both renewable and fossil‐based energy sources. Hence, the environmental impact of such networks is studied. Lastly, the proposed CHAMEN model is extended to handle multi‐objective optimization (MOO) of both environmental impact and economics, to identify an optimum network configuration.

A trust region framework is presented to synthesize heat exchanger network with detailed exchanger designs. The heat exchanger network (HEN) is first synthesized using the stage wise superstructure (SWS) formulation of Yee and Grossmann (1990). After a topology is found in this step the heat exchangers and the connections, flows and intermediate temperatures are designed, using the first principles based differential algebraic (DAE) model presented in our previous work. These detailed DAEs for heat exchanger design are incorporated within a nonlinear programming (NLP) model using reduced order models and solved using NLP solver IPOPT with a trust region method. The results show that the new method is faster than the previous approaches while providing comparable results.

Developing a flexible heat exchanger network, that will remain operable in the face of potential variations in stream parameters, or variables, around some nominal values, especially for large problems, are difficult to solve simultaneously. A hybrid synthesis approach that systematically combines sequential and mathematical programming techniques may be more suitable to adopt. The proposed method presented in this paper entails firstly generating two representative single period networks using the Pinch Technology Stream Temperature versus Enthalpy Plot (STEP) approach and the multi-period stage-wise superstructure model for heat exchanger network synthesis. Streams for the representative network are obtained using the largest stream heat demand as criteria. The second stage of the proposed method entails generating a reduced multi-period stage-wise superstructure, using a combination of the matches obtained in the single period representative networks of the first step as initialising matches. The solution of the reduced superstructure of the second step, which is solved as a Mixed Integer Non-Linear Programming (MINLP) model, is then selected as the best multi-period network. The newly developed method of this paper is tested using an example from the literature. One of the solutions obtained is only 0.9 % higher, in terms of total annual cost, than the best solution presented in the literature but has the benefit of a fewer number of units.

This work provides the first systematic critical review of mass exchanger network synthesis literature. Mass exchanger networks play a central role in many pollution reduction and resource utilisation processes and contain many complex decisions to be made including exchanger types, sizing, and mass separating agent selection. We present a comprehensive review of the key milestones in the development of methods for mass exchanger network synthesis and focus on the key challenges that have hindered research in this area from flourishing in the manner of the conceptually similar heat exchanger network synthesis problem. We find that several important research questions remain for the methods to find wider use in industry. More efficient techniques for solving nonconvex mixed-integer nonlinear programs and better methods of including more accurate, higher-order unit models for industrial problems within network optimisation problems are particularly important, as current methods provide highly simplified unit representations that do not take into account many important practical design considerations that have significant cost implications. Furthermore, we identify significant potential for further research into increasing the scope of the problem to include issues such as flexibility and controllability, inter-plant mass exchanger networks, batch processes, retrofit and further integration of heat and mass exchanger networks, with research into these domains limited. Through further research of these under-developed applications of mass exchanger network synthesis, we envision that techniques for mass integration can become a powerful tool to enhance mass integration techniques for sustainable cleaner production technology. •Existing synthesis methods have been mainly graphical and deterministic.•More robust methods should be developed for multicomponent systems.•Most methods have included simplified unit designs rather than detailed ones.•Controllability and flexibility despite being crucial have received less attention.•Opportunities for interplant mass integration should be explored.

This paper presents KIPET (Kinetic Parameter Estimation Toolkit) an open-source toolbox for the determination of kinetic parameters from a variety of experimental datasets including spectra and concentrations. KIPET seeks to overcome limitations of standard parameter estimation packages by applying a unified optimization framework based on maximum likelihood principles and large-scale nonlinear programming strategies for solving estimation problems that involve systems of nonlinear differential algebraic equations (DAEs). The software is based on recent advances proposed by Chen et al. (2016) and puts their original framework into an accessible framework for practitioners and academics. The software package includes tools for data preprocessing, estimability analysis, and determination of parameter confidence levels for a variety of problem types. In addition KIPET introduces informative wavelength selection to improve the lack of fit. All these features have been implemented in Python with the algebraic modeling package Pyomo. KIPET exploits the flexibility of Pyomo to formulate and discretize the dynamic optimization problems that arise in the parameter estimation algorithms. The solution of the optimization problems is obtained with the nonlinear solver IPOPT and confidence intervals are obtained through the use of either sIPOPT or a newly developed tool, k_aug. The capabilities as well as ease of use of KIPET are demonstrated with a number of examples.