海军航空大学学报2025，Vol.40Issue(5)：669-684,16.DOI:10.7682/j.issn.2097-1427.2025.05.003
模型预测直接推力控制的多目标参数整定方法
Multi-Objective Parameter Tuning Method for Model Predictive Direct Thrust Control
摘要
Abstract
Model Predictive Control(MPC)is capable of achieving command tracking while ensuring that all con-straints remain within limits,which is why it has been applied to Direct Thrust Control(DTC)for gas turbine engines.However,MPC involves numerous tuning parameters that significantly affect DTC performance,making it extremely necessary to develop effective parameter tuning methods.To address this issue,an MPC-based DTC architecture is de-signed and the impact of the tuning parameters is analyzed.Based on this analysis,a parameter tuning method based on a Multi-Objective Grey Wolf Optimizer(MOGWO)is proposed.The proposed method formulates corresponding multi-objective optimization index functions by integrating the objectives of DTC and employs MOGWO to solve for the Pa-reto front in order to obtain the desired tuning parameters.Hardware-in-the-loop(HIL)experimental results demonstrate that the proposed MPC-based DTC architecture reduces the settling time by 0.13 s compared with the baseline.More-over,the proposed MOGWO-based parameter tuning method better avoids local optima than the control group and is able to obtain tuning parameters from both rapidity-optimal and stability-optimal perspectives.Finally,experiments un-der component degradation conditions verify the robustness of the proposed control architecture and parameter tuning method.关键词
直接推力控制/模型预测控制/整定/多目标优化/多目标灰狼算法Key words
direct thrust control/model predictive control/tuning/multi-objective optimization/multi-objective grey wolf optimizer分类
航空航天引用本文复制引用
王根常,刘晓锋..模型预测直接推力控制的多目标参数整定方法[J].海军航空大学学报,2025,40(5):669-684,16.基金项目
国家自然科学基金(52372371) （52372371）
