Abstract
A key factor for energy-efficient industrial clusters is the recovery of waste heat. To this end, thermal energy storage (TES) is an appealing technology that facilitates dynamic heat integration between supplier and consumer plants. A long-term strategy for energy savings must involve adequate consideration for the optimal design of the TES. From an industrial perspective, finding the capacity of the TES unit is often based on heuristic rules which may lead to suboptimal design. This approach does not account for the short-term variability in operation of the TES system. Scenario-based stochastic programming approaches, where the operational uncertainty is described in form of discrete scenarios, can be used to find the best design for the TES system. We present two problem formulations for finding the optimal capacity of the TES unit. The first is a single-level formulation where the design and operating constraints are combined for all scenarios, with the objective of minimizing the combined cost of design and operation. The second is a bilevel formulation where the design decisions are taken on the upper level to minimize overall system cost, whereas the lower level problems (one per scenario) represent the optimal operation for the chosen design variables, each minimizing the operating cost for their respective scenarios. We compare the results of the two approaches with an illustrative case study of an industrial cluster with one supplier plant and one consumer plant exchanging heat via a TES unit. © 2020 Elsevier B.V.
