摘要
针对半桥谐振逆变型磁共振式无线充电(MCR-WCT)系统工作状态模糊、高阶非线性、控制理论不成熟等问题,在建立半桥谐振逆变电路等效模型的基础上,采用广义状态空间平均法(GSSA)对MCR-WCT系统进行大信号和小信号建模,在GSSA模型的基础上设计电流单闭环控制器,制定基于BP神经网络的自整定PID控制策略。最后,通过Matlab编程对GSSA大信号模型进行暂态和稳态分析,对比Simulink模型仿真结果验证GSSA模型的可行性;通过Matlab仿真对比经典PID控制和BP神经网络自整定PID控制策略,在电流设定值为1 A的阶跃响应中,BP神经网络自整定PID控制在0.25 ms内达到稳态,稳态误差在2%内,最大超调量只有5%,相比经典PID控制具有更好的动静态性能。
In order to solve the problems in the half-bridge inverter based magnetic coupling resonant wireless charging(MCR-WCT)system,including working status ambiguity,high order nonlinearity and immature control theory,the equivalent model of half-bridge resonant inverter was established.And the generalized state-space averaging method(GSSA)was used to build up the large and small signal model of the MCR-WCT system.Furthermore,the current single closed loop controller was designed on the basis of the GSSA model,and then the self-tuning PID control strategy based on BP neural network came up.Finally,the transient and steady-state analysis of the GSSA large signal model was carried out through Matlab simulation,and the simulation results of the Simulink model were compared to verify the feasibility of the GSSA model.The traditional PID control algorithm and the BP neural network self-tuning PID control algorithm were compared through Matlab simulation.In the step response of the current setting value of 1 A,the BP neural network selftuning PID control algorithm reached the steady-state within 0.25 ms,while the steady-state error was within 2%,and the overdose was only 5%.Compared with the traditional PID control,it had better dynamic and static performance.
作者
曾得志
薛家祥
ZENG Dezhi;XUE Jiaxiang(School of Mechanical and Automotive Engineering,South China University of Technology,Guangzhou 510640,China)
出处
《中国测试》
CAS
北大核心
2020年第2期110-116,154,共8页
China Measurement & Test
基金
福建省自然基金项目(2018J01541)
广州市南沙区科技计划项目(2017CX009)
2015东莞市引进第三批创新科研团队项目(2017360004004)
关键词
磁共振式无线充电技术
半桥谐振逆变
广义状态空间平均法
BP神经网络自整定PID控制
magnetic coupling resonant wireless charging technology
half-bridge resonant inverter
generalized state-space averaging
BP neural network self-tuning PID control
