A Novel Approach for Linearization of a MINLP Stage-Wise Superstructure Formulation

Beck, Anton; Hofmann, René

doi:10.1016/j.compchemeng.2018.01.010

Cited by 23 publications

(16 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…E ee e e ee e e 1 1 1, (34) where τ is a multiplier of the order of TAC, UC s is the unit cost of utility s ($/kW), and γ is the capital annualization factor. This completes our MINLP formulation (eqs 1−8 and 10−34) for minimizing COST with all variables being non-negative.…”

Section: Industrial and Engineering Chemistry Researchmentioning

confidence: 99%

“…Pavão et al enhanced SS by introducing utility exchangers in parallel in each stage. While Beck and Hofmann , presented tightening strategies and a novel linearization for SS, Huang and Karimi found better solutions by combining the hyperstructure of Floudas et al with the SS of Yee and Grossmann . In another vein, Jongsuwat et al introduced substages at each stage to overcome the limitations of SS and used an ad-hoc solution strategy.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Unified Heat Exchanger Network Synthesis via a Stageless Superstructure

Nair

Karimi

2019

Ind. Eng. Chem. Res.

View full text Add to dashboard Cite

The existing mathematical programming-based methods for simultaneous heat exchanger network synthesis (HENS) have used either match-centric or stage-centric superstructures. Most works over the past four decades have primarily used various modifications of essentially two superstructures whose configurational limitations are well established. We revisit a novel, promising, but unsuccessful, exchanger-centric superstructure proposed by Huang and Karimi (Chem. Eng. Sci.201398231245). We modify the superstructure and the associated mixed-integer nonlinear programming formulation substantially, and develop an efficient algorithm for its solution. The superstructure simply assumes a pool of two-stream exchangers, to which hot and cold streams are assigned. Given a sufficiently large pool, it embeds all possible heat exchanger network (HEN) configurations (in contrast to previous superstructures) including repeated matches, cross flows, bypasses, etc., and allows multiple utilities. The novel HEN configurations obtained for three case studies and improvements in their total annual costs highlight the advantages of the proposed superstructure and model.

show abstract

Section: Industrial and Engineering Chemistry Researchmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Unified Heat Exchanger Network Synthesis via a Stageless Superstructure

Nair

Karimi

2019

Ind. Eng. Chem. Res.

View full text Add to dashboard Cite

show abstract

“…Numerous design technologies and methods have been developed with extensive application to industrial practice, as stated in review articles 1,2 and research publications, 3,4 to name a few. Of all the previous works on HEN synthesis, the approach that gains dominance is the one where a superstructure is proposed together with the formulation of a mixed‐integer nonlinear model (MINLM), which can be solved using mixed‐integer nonlinear programming (MINLP) procedures, 5,6 many times using decompositions, 7,8 or stochastic‐based methods (metaheuristics‐based methods), such as genetic algorithms, 9,10 simulated algorithms, 11,12 particle swarm optimization, 13,14 hybridization between different algorithms, 15,16 etc. We do not elaborate further on the literature review of the stochastic‐based methods, which cannot guarantee optimality, much less global optimality, and in general need specialized parameter tuning for a good computational performance.…”

Section: Introductionmentioning

confidence: 99%

Globally optimal synthesis of heat exchanger networks. Part III: Non‐isothermal mixing in minimal and non‐minimal networks

et al. 2021

View full text Add to dashboard Cite

In this work, the enumeration algorithms presented in Parts I and II for the globally optimal synthesis of minimal and non‐minimal heat exchanger networks are extended to consider non‐isothermal mixing. New mathematical models, including non‐isothermal mixing constraints, are proposed to target the bounds of energy consumption and the binding exchanger minimum approximation temperature. These models are solved using the algorithms, which involve solving systems of equations instead of mathematical programming. Three global optimization strategies are proposed to optimize each enumerated structure, involving the use of a global solver directly, or the use of a Golden Search based on energy consumption and a flowrate optimization model considering non‐isothermal mixing. The flowrate optimization model is reformulated as a convex problem, which is solved by using nonlinear programming or a mathematical programming‐free methodology, that is, solving Karush–Kuhn–Tucker equations. A new Global Optimum Search Algorithm is developed and examples are tested comparing different optimization strategies.

show abstract

“…The two models were mostly used in heat exchange networks, which is a research system from the perspective of energy flow. [ 9,10 ] In addition to the optimized design of the process, the mass exchange network and the heat exchange network are inseparable. The hybrid network of mass exchange and heat exchange is becoming a new field of process comprehensive research.…”

Section: Introductionmentioning

confidence: 99%

An ontology‐based procedure knowledge framework for the process industry

Cao

Zhu

2020

Can J Chem Eng

View full text Add to dashboard Cite

Process industry enterprises rely heavily on expert experience in production, and expert knowledge is stored in multiple formats like, pictures, texts, videos, etc. Several isolated islands of information or knowledge are formed, and they are difficult to store, share, and expand upon. Therefore, the same procedure appears multiple times in different application scenarios, such as optimization scheduling, optimization operation, or fault diagnosis. If redundant knowledge is paid too much attention, decision-makers will not be able to comprehensively consider business knowledge. Thus, a novel framework is proposed for the process industry to manage procedure knowledge. This paper first divides the detailed procedure of process plants into four layers, that is, raw materials, intermediate materials, operation, and products based on the P-graph theory, which involves virtual hierarchies, nodes, and operations. Second, the domain ontology-based procedure knowledge model is constructed. Finally, two experimental cases, the Tennessee Eastman (TE) and the ethylene production process, are studied to verify the procedure knowledge framework (PKF). The PKF provides a useful foundation for the subsequent construction of a superstructure model based on the P-graph theory and is different from the previous industrial process industrial ontology model. The PKF has a theoretical basis for simulation calculation and makes future work based on PKF more accurate and interpretable.

show abstract

A Novel Approach for Linearization of a MINLP Stage-Wise Superstructure Formulation

Cited by 23 publications

References 18 publications

Unified Heat Exchanger Network Synthesis via a Stageless Superstructure

Unified Heat Exchanger Network Synthesis via a Stageless Superstructure

Globally optimal synthesis of heat exchanger networks. Part III: Non‐isothermal mixing in minimal and non‐minimal networks

An ontology‐based procedure knowledge framework for the process industry

Contact Info

Product

Resources

About