航空发动机神经网络自学习PID控制

推进技术 ›› 2007, Vol. 28 ›› Issue (3): 313-316.

航空发动机神经网络自学习PID控制

姚华,袁鸯,鲍亮亮,孙健国

南京航空航天大学能源与动力学院江苏南京210016;中国一航动力控制系统研究所;江苏南京210016;南京航空航天大学能源与动力学院江苏南京210016;南京航空航天大学能源与动力学院江苏南京210017;南京航空航天大学能源与动力学院江苏南京210018

发布日期:2021-08-15
基金资助:
国家自然科学基金(50576033);航空科学基金(04C52019)资助项目

Self-learning PID control based on neural networks for aeroengines

Coll.of Energy and Power,Nanjing Univ.of Aeronautics and Astronautics,Nanjing 210016,China;China Aviation Motor Control System Inst.,Wuxi 214063,China;Coll.of Energy and Power,Nanjing Univ.of Aeronautics and Astronautics,Nanjing 210016,China;Coll.of Energy and Power,Nanjing Univ.of Aeronautics and Astronautics,Nanjing 210017,China;Coll.of Energy and Power,Nanjing Univ.of Aeronautics and Astronautics,Nanjing 210018,China

Published:2021-08-15

摘要/Abstract

摘要： 将神经网络与传统的PID控制相结合,构成神经网络自学习PID控制,用神经网络在线整定PID控制器的比例、积分及微分三个参数,使被控对象跟踪理想参考模型的输出。该系统具有自学习能力,能适用于非线性、时变的被控对象。将神经网络自学习PID控制方法用于航空发动机全包线控制以及蜕化发动机的控制,进行了数字仿真,验证了该方法的有效性。

关键词: 航空、航天推进系统;神经网络;PID控制;自学习+

Abstract: A self-learning PID controller based on neural networks and conventional PID control was developed.The parameters of PID controller are tuned on-line with the neural networks to make the output of the controlled plant follow the desired output of a reference model.The resulting control system is capable of self-learning and can be used for controlling nonlinear and time-varying plant.The proposed method is applied to aeroengine control.Digital simulation results show that the self-learning PID control proposed is effective to control nominal and deteriorated aeroengine in full envelope.

Key words: Aerospace propulsion system;Neural networks;PID control;Self-learning+

姚华,袁鸯,鲍亮亮,孙健国. 航空发动机神经网络自学习PID控制[J]. 推进技术, 2007, 28(3): 313-316.