Optimization scheduling of ship integrated energy system with organic rankine cycle waste heat recovery based on SCSSA-CNNBiLSTM load forecasting

AbstractThis paper proposes a new Ship Integrated Energy System with Organic Rankine Cycle waste heat recovery (SIES-ORC) considering the multi-energy load demand of ships. The multi-energy load forecasting method, Sine-cosine and Cauchy mutation Sparrow Search Algorithm - Convolutional Neural Networks and Bidirectional Long and Short-Term Memory (SCSSA-CNNBiLSTM), is applied to achieve the optimal schedule of the designed SIES-ORC. Based on the predicted multi-energy load data, dynamic operation and optimization scheduling are implemented for a SIES including carbon capture, life-cycle emission accounting of fuels, and carbon trading, and a ship operation optimization strategy are obtained. The results show that the significant advantages are in four types of load forecasting, with mean absolute percentage errors for mechanical, electrical, heat, and cooling energy predictions being 27.71%, 0.23%, 1.11%, and 0.33%, respectively. The initial investment in the system has the highest proportion and the carbon emissions are mainly produced by the Dual Fuel Engine (DFE), the Dual Fuel Generator Sets (DFGS), and the Fuel Cell (FC), with the DFGS accounting for 67.86%, the DFE accounting for 32.14%, and the FC being negligible. This study provides effective technical support for the accurate prediction of complex multi-energy loads in integrated energy systems.