Ship-board integrated power systems (IPS) play a crucial role in the all-electric ships (AES) initiative, where the power generation, conversion and distribution are done on a common electric platform. Along with many advantages of the IPS, its inter-connectivity presents a major opportunity as well a challenge for reconfiguration control. To ensure overall optimal performance while sustaining the critical ship service loads, the reconfiguration problem can be formulated as a trajectory optimization problem with the inputs to power sources as optimization variables. The high dimensional and nonlinear nature, along with the extended time horizon involved, makes many grid and gradient based methodologies infeasible for this application, especially when real-time requirements are considered. This paper, motivated by the IPS reconfiguration control, proposes strategies for improving the real-time computational efficiency of the IPS optimization by leveraging the time scale separation of its dynamics. The trade-offs between the computational efficiency and optimization accuracy are analyzed and a numerical example is provided to illustrate the effectiveness of the proposed algorithm.
A scalable time–space multiscale domain decomposition method: adaptive time scale separation