Multiperiod work and heat integration

• Published in: Energy conversion and management 2021-01, Vol.227, p.113587, Article 113587
• Main Authors: Pavão, Leandro V., Miranda, Camila B., Caballero, Jose A., Ravagnani, Mauro A.S.S., Costa, Caliane B.B.
• Format: Article
• Language: English
• Subjects: Heat; Heat exchange; Heat exchangers; Heat recovery systems; Heat transfer; Heuristic methods; Integration; Meta-heuristics; Multiperiod work and heat exchange networks; Network synthesis; Optimization; Process synthesis; Work and heat integration
• ISSN: 0196-8904; 1879-2227
• DOI: 10.1016/j.enconman.2020.113587

•A method for multiperiod work and heat integration is developed.•Steps comprise the application of optimization models solved by meta-heuristics.•Nominal and critical (non-nominal) conditions are defined as periods.•Cases with seven and fourteen periods are tackled.•Solutions with reasonable additional costs and efficient units usage are obtained. The synthesis of multiperiod heat exchanger networks (HEN) is a well-studied topic in heat integration. Several methods for identifying heat exchanger network designs that are able to feasibly operate under multiple conditions have been presented. Multiperiod models are certainly a notable form of achieving such resilient designs. In work and heat integration, however, solutions presented so far are for nominal conditions only. This work presents a step-wise optimization-based multiperiod work and heat exchange network synthesis framework. Hybrid meta-heuristic methods are used in the optimization steps. The methodology is able to obtain work and heat exchanger networks (WHENs) that are able to operate under multiple known scenarios. A set of critical conditions for stream properties in work integration is proposed. When these scenarios are modeled as finite operating periods (which are here referred to as non-nominal periods), a WHEN which can feasibly operate under nominal and critical conditions can be obtained. An example is tackled in two cases: the first, with one nominal and six critical, non-nominal periods; the second with two nominal and twelve non-nominal periods. Note that with that number of periods, the problem is considerably more complex than in multiperiod HEN synthesis (which usually comprises three or four periods). Solutions obtained with the present method are compared to those obtaining by simply merging single-period solutions obtained for each period individually. Capital investments are 30.2% and 58.2% lower in Cases 1 and 2 than in straightforwardly merged solutions. The capacity utilization parameters also demonstrate that the overdesign issue is notably reduced in these solutions.