Energy-efficient dispatch of multiple-chiller systems using hybrid exchange market and genetic algorithm

Abstract

The inefficient dispatch of multiple-chiller systems results in drastic energy consumption, which should be investigated by optimal allocating of cooling loads between chillers in a building. Due to the nonlinearity of the dispatching problem, heuristic algorithms are more prevalent in this field. In this paper, a hybrid exchange market and genetic algorithm (EMGA) is established to solve the optimal chiller loading (OCL) problem. The decision variables of the problem are the partial load ratio (PLR) of chillers determining the power consumption of each chiller unit, and the objective of the OCL problem is to minimize the total power consumption of the multiple-chiller system. The effectiveness of the proposed algorithm is verified by a fair comparison of the results with similar attributes in three identical cases. It is shown that the proposed EMGA outperforms comparable algorithms in terms of convergence speed and finding optimal solutions. Moreover, the problem is modeled as mixed-integer non-linear programming (MINLP) in the general algebraic modeling system (GAMS) to solve mathematically using different solvers. It was shown that the mathematical approaches adopted by GAMS have a high probability of trapping in local optima when the decision variables are increasing, whiles the optimum solutions by the proposed EMGA are superior.