项目名称: 大惯量非线性系统的多驱动控制
项目编号: No.61273150
项目类型: 面上项目
立项/批准年度: 2013
项目学科: 自动化技术、计算机技术
项目作者: 任雪梅
作者单位: 北京理工大学
项目金额: 80万元
中文摘要: 项目从大惯量非线性系统精确控制的需求出发,针对多电机同步驱动控制中存在的关键问题,开展偏置力矩优化设计、多电机同步控制、机械谐振抑制、非光滑动态补偿控制研究。研究内容包括: 提出了偏置力矩的优化调节算法,以最小能耗解决齿隙不可控问题;给出主从驱动子系统的非线性速度差反馈控制,实现多电机快速同步控制;为了克服大惯量系统的机械谐振和抑制外界干扰,研究能同时估计负载加速度和干扰的有限时间扩展观测器,给出了基于扩展观测器的快速递归动态面控制,以期望动态在有限时间内达到控制要求;针对控制系统中存在的分段函数,提出了分段神经网络模型,研究了摩擦特性的分段模型并建立神经网络补偿控制,提高系统的低速性能;建立大惯量非线性系统多驱动控制的整体模型及综合控制方案,将理论研究成果应用到实际四电机同步控制系统。项目研究成果为研发高精度大惯量系统的多驱动控制提供理论支持,为解决大惯量非线性系统的控制提供新方法。
中文关键词: 大惯量非线性系统;多驱动控制;系统辨识;神经网络;自适应控制
英文摘要: From the demands of accurate control for large inertia nonlinear systems, and aiming at solving the key problems of the multi-motor drive control, this project investigates the optimization design of bias torque, multi-motor synchronization control, mechanical resonance rejection and compensation control for non-smooth dynamics. The project includes the following contents.The optimization update scheme of the bias torque is proposed to solve the uncontrollable problem of backlash with minimum energy. The nonlinear speed-deviation feedback control is established for master-slave drive systems to realize the fast synchronization control. The load acceleration and disturbances are estimated by the finite time extended observer, and the recursive dynamic surface control based on the extended observer is given to solve the mechanical resonance which can achieve the control performances in the limited time. The piecewise neural networks are proposed for the piecewise functions in the control systems and the corresponding identification method for the friction is given. The compensation control based on the neural networks is proposed to improve the low speed performances. The overall identification model and control scheme are established for large inertia nonlinear systems with multi-drive control.The theoretical res
英文关键词: Large inertia nonlinear systems;Multi-drive control;System identification;Neural networks;Adaptive control