中国电机工程学报2026，Vol.46Issue(5)：2007-2021,中插22,16.DOI:10.13334/j.0258-8013.pcsee.242824
考虑恒功率负载的直流微电网控制策略:基于固定时间前馈解耦的自主无源模型预测技术
Considering Constant Power Load for DC Microgrid Control Strategy:Fixed-time Feedforward Decoupling-based Autonomous Passivity Model Prediction Technique
摘要
Abstract
Considering that DC microgrids should exhibit excellent robustness and coordination when facing source-load side disturbances and the instability problem caused by constant power load(CPL),the paper proposes an autonomous passivity model predictive control strategy based on fixed-time feedforward decoupling for boost converter parallel system as the object of study.Firstly,the fixed-time sliding mode disturbance observer(FTSMDO)with online estimation of source-load side information is designed,which not only reduces number of system sensors,but also avoids the coupling disturbance of coordinate transformation.Then,combining passivity and model predictive control theories to propose the advanced control strategy that possess the advantages of both and compensate their respective weaknesses,while realizing decentralized power sharing autonomous control structure under fixed-time feed-forward decoupling of observers.Moreover,the stability analysis method of the bifurcation diagram based on the discrete-time model of the system is proposed,which explores the critical power of constant power load and the selection criterion of decoupling control parameters.Finally,the effectiveness and superiority of proposed controller is verified through simulation and experimental results with less system prior knowledge.关键词
直流微电网/恒功率负载/固定时间滑模扰动观测器/自主无源模型预测控制/稳定性分析Key words
DC microgrid/constant power load(CPL)/autonomous passivity model predictive control(APMPC)/fixed-time sliding mode disturbance observer(FTSMDO)/stability analysis分类
信息技术与安全科学引用本文复制引用
张泽华,王盼宝,刘涵颖,周洋,王卫,徐殿国..考虑恒功率负载的直流微电网控制策略:基于固定时间前馈解耦的自主无源模型预测技术[J].中国电机工程学报,2026,46(5):2007-2021,中插22,16.基金项目
国家自然科学基金项目(52377173).Project Supported by National Natural Science Foundation of China(52377173). （52377173）
