异步电动机调速系统自适应辨识的CMAC-ADRC算法

发布时间：2018-06-18 13:49

本文选题：异步电动机 + 小脑模型神经网络　；参考：《农业机械学报》2015年03期

【摘要】：针对异步电动机调速系统快速响应时启动超调量大的问题,提出了一种基于自适应参数辨识的小脑模型神经网络复合自抗扰控制(CMAC-ADRC)的控制算法。将CMAC与ADRC各自的优点相结合,利用CMAC神经网络实现前馈控制,通过在线学习来抑制系统的超调量,增强系统的鲁棒性能,提高系统的快速性能,利用ADRC技术实现反馈控制,进一步增强系统的抗干扰能力。利用参考模型自适应参数辨识技术对转动惯量进行辨识,优化自抗扰补偿系数。以变频器结合异步电动机为控制对象,进行仿真,基于自适应参数辨识的CMAC-ADRC控制算法的干扰响应幅度是一阶优化自抗扰控制下干扰响应幅度的44.57%,是小脑模型神经网络复合比例-微分(CMACPD)控制下干扰响应幅度的17.69%,干扰恢复时间是一阶优化自抗扰控制下干扰恢复时间的50%,是CMAC-PD控制下恢复时间的60%。搭建MCU-CPLD-DSP控制平台进行了实验,基于自适应参数辨识的CMAC-ADRC控制算法的超调量是一阶优化自抗扰控制的45.49%,上升时间是一阶优化自抗扰控制的53.33%,干扰响应幅度是一阶优化自抗扰控制干扰响应幅度的71%,干扰恢复时间是一阶优化自抗扰控制干扰恢复时间的76.47%。
[Abstract]:In order to solve the problem of large overshoot in the fast response of asynchronous motor speed control system, a control algorithm of CMAC-ADRC based on adaptive parameter identification for cerebellar model neural network combined with active disturbance rejection control (ADRC) is proposed. Combining the advantages of CMAC and ADRC, the feedforward control is realized by using CMAC neural network. Through on-line learning, the overshoot of the system is restrained, the robustness of the system is enhanced, and the fast performance of the system is improved. The feedback control is realized by ADRC technology. Further enhance the anti-interference ability of the system. The adaptive parameter identification technique of reference model is used to identify the moment of inertia and to optimize the compensation coefficient of auto disturbance rejection. The inverter is combined with the asynchronous motor as the control object, and the simulation is carried out. The interference response amplitude of the CMAC-ADRC control algorithm based on adaptive parameter identification is 44.57 of the interference response amplitude under the first-order optimization auto disturbance rejection control, and is 17.69g of the interference response amplitude under the control of the cerebellar model neural network compound proportion-differential CMACPD. The disturbance recovery time is 50 parts of the first order optimal disturbance recovery time under active disturbance rejection control, which is 60% of the CMAC-PD control. The MCU-CPLD-DSP control platform is built. The overshoot of the CMAC-ADRC control algorithm based on adaptive parameter identification is 45.49 of the first-order optimal ADRC, the rising time is 53.33 of the first-order optimal ADRC, and the amplitude of the disturbance response is the first-order optimal ADRC interference response amplitude. The interference recovery time is 76.4747 of the first order optimal disturbance rejection control interference recovery time.
【作者单位】：江苏大学电气信息工程学院;淮阴工学院电子与电气工程学院;机械工业设施农业测控技术与装备重点实验室;
【基金】：国家自然科学基金资助项目(60874014、51273154) 江苏高校优势学科建设工程资助项目(PAPD,NO.6-2011) 江苏省2013年度普通高校研究生科研创新计划资助项目(CXLX13_669) 江苏省科技支撑计划资助项目(BE2013402) 淮安市农业科技指导性项目(HANZ2014007)
【分类号】：TM343

【参考文献】