Multiple optimal solutions in synthesis, design and operation optimization of a marine energy system with the differential evolution algorithm

• Obtaining multiple solutions in synthesis design and operation optimization problems. • Efficient algorithm to handle components selection. • Multiple configurations investigated (with and without waste heat recovery) with detailed analysis of the results. This article deals with the optimization of a ship energy system on multiple levels (synthesis, design and operation). These complex problems often induce many local optima, making it difficult to obtain reliable optimization results. A clustering method based on the absence or presence of components in the architecture coupled with an evolutionary algorithm (i.e Differential Evolution) is proposed to tackle this issue. The method enables multiple optimal solutions to be identified for a real-world optimization problem. The reference ship for this study is a destroyer but the use of the Admiralty coefficient in the model description makes the algorithm easily adaptable for any kind of ships. The specific fuel consumption is pre-calculated for groups of components to alleviate the computational cost of the optimization problem. Two numerical cases are computed representing the ship energy system with or without a heat recovery and a heat creation system to illustrate the capabilities of the method. The objective in both cases is to minimize the weight of the ship (fuel consumption + components weights) for a given mission profile. The influence of the clustering technique, population size and repeatability of the differential algorithm is also investigated to assess the reliability of the method.