Hybrid Model Linearization Predictive Control for DC Bus Stabilization in Spacecraft Combined Power System

Abstract

In spacecraft, facing different loading conditions of subsystems, especially of permanent magnet synchronous motor (PMSM) system, the bus voltage has large fluctuation, which will seriously affect the power quality of the spacecraft combined power system. Therefore, it is necessary to introduce a bidirectional DC–DC converter to suppress the fluctuation of the bus voltage. Aiming at the need of controlling the bidirectional DC–DC converter, a hybrid model linearization predictive control (HMLPC) algorithm is proposed. According to the strong nonlinear hybrid characteristics of the bidirectional DC–DC converter, an optimal linear approximation model at the sampling point is established using the Jacobian matrix, based on which a piecewise affine (PWA) model is established for Buck and Boost mode, and then, a hybrid linearization predictive model is constructed. Ultimately, by introducing a two-period PWA and multistep prediction strategy, proposing a prediction error, and solving the constrained finite time optimization control (CFTOC) problem online, an optimal control law is derived. In addition, a stability analysis is provided for the proposed control algorithm. In this article, a simulative validation and a comparative analysis are presented; moreover, a combined power test platform is built to validate the proposed control algorithm. Both simulation and experiment results show that the HMLPC algorithm can effectively suppress the fluctuation of bus voltage and enhance the anti-interference ability of the system.