Untitled

The minimum capital cost/total annualized cost (MCC/MTAC) problem is investigated, within the infinite-dimensional state-space (IDEAS) framework, for homogeneous and isothermal reactor networks. The resulting mathematical formulation is an infinite-dimensional program with linear constraints and a separable concave objective function to be minimized. The global optimum of this infinite program is approximated through global solution of a series of finite-dimensional programs with concave separable objective functions and linear feasible regions. A branch-and-bound algorithm is used to solve globally each of these finite programs. The proposed methodology is illustrated with a reactor network synthesis case study aiming at capital cost minimization.

Section: Introductionmentioning

confidence: 88%

Global Capital/Total Annualized Cost Minimization of Homogeneous and Isothermal Reactor Networks

Zhou¹,

Manousiouthakis²

2008

“…Considering the characteristic of the CONOPT solver for NLP problems, a good initial point from the reduced model is strongly recommended, which can be obtained through the method proposed by Esposito and Floudas. 48 The model is reformulated into a MINLP, which belongs to NP-Hard problems. CONOPT cannot guarantee global optimality.…”

Section: Modeling and Solution Strategymentioning

confidence: 99%

“…The complementarities and related equations are introduced as follows: so that the model is formulated as a relaxed MINLP model where much fewer NLP subproblems need to be solved. Considering the characteristic of the CONOPT solver for NLP problems, a good initial point from the reduced model is strongly recommended, which can be obtained through the method proposed by Esposito and Floudas . The model is reformulated into a MINLP, which belongs to NP-Hard problems.…”

Section: Modeling and Solution Strategymentioning

confidence: 99%

Synthesis and Optimization of Reaction–Separation–Recycle Systems with Complex Distillation Sequences

Zou

Wei-xuan

et al. 2020

In this work, a novel synthesis superstructure for reaction−separation−recycle systems with complex distillation sequences is proposed. In the superstructure, complete interconnections among reactor modules, as well as necessary connections between reactor network and distillation sequence, are considered. State Task Network representation is used to describe the interconnections among distillation columns, so that all possible single-column sequences including sloppy separation ones are generated conveniently. Based on this, component distribution in distillation columns through Underwood equations is introduced to provide products with enough purity and multiple component mixtures as feedback to the reactor network to achieve the optimal composition distribution in the total flowsheet. Furthermore, to correctly estimate the number of theoretical stages of sloppy separation, the Hengstebeck−Underwood−Gilliland correlation method is employed, replacing the conventional Fenske− Underwood−Gilliland method in the distillation column design submodel. Finally, three literature examples are performed to demonstrate the feasibility and validity of the proposed methodology.

“…The majority of the work in the literature, however, addresses only the problem of continuous-flow reactor networks. After selecting the continuous-flow operating mode, the attainable region method, − superstructure optimization, − stochastic optimization, , and dynamic optimization − have been used to arrive at the optimal reactor network and/or the optimal process. Note that some of the referenced literature merges several of the aforementioned categories and may focus only on the reactor or on the full process synthesis problem.…”

Section: Introductionmentioning

confidence: 99%

A Unified Reactor Network Synthesis Framework for Simultaneous Consideration of Batch and Continuous-Flow Reactor Alternatives

Costandy

Edgar

Bâldea

2021

Determining the optimal operating mode (batch or continuous-flow) for chemical manufacturing remains a pressing open problem. Focusing on reactors, an accurate determination must account for fundamental differences between operating modes, the size and cost of the respective reactor units, and the potential of utilizing reactor networks to optimize performance. We present a reactor network synthesis framework that simultaneously considers batch and continuous-flow options while accounting for all the aforementioned factors. We introduce a non-dimensional model structure that encompasses all operating mode options, and its discretized representation, the orthogonal collocation on finite elements for reactors (OCFERE). This allows us to formulate a unified economic model and a framework for optimal reactor network synthesis. An illustrative case study is also provided.