Multi-objective coordinated scheduling of locomotive depot microgrids based on spatiotemporal-energy decoupling and virtual energy storage

To address transient grid impacts and high operating costs caused by high-frequency, high-power charging of new-energy locomotives in locomotive depot microgrids, this paper innovatively proposes a timetable-driven spatiotemporal-energy decoupling collaborative scheduling method. First, a locomotive spatiotemporal state correlation matrix and a discrete energy mapping model are established, reconstructing the rigid traffic load into a virtual energy storage system with great large-capacity temporal translation potential. Second, a multi-objective optimal scheduling model is constructed to minimize microgrid comprehensive operation and maintenance costs and main grid interactive power fluctuation, considering locomotive smooth ramping and battery safe operation boundaries. For the high-dimensional, discontinuous mixed-integer nonlinear programming problem, an improved multi-objective particle swarm optimization–mantis shrimp optimization algorithm integrated with physical constraint masking is proposed, which reduces dimension via the underlying temporal feasible region filtering mechanism and achieves efficient coordination between global exploration and local exploitation. Multi-scenario simulations show the strategy guides locomotive clusters to form refined temporal peak-shifting. Compared with unordered charging, system comprehensive operation and maintenance costs decrease by 59.6%, main grid interactive power fluctuation by 85.2%, achieving high economic benefits, new-energy local consumption rate and grid-connection friendliness while ensuring transportation rigid demand.