【作者】 魏熙乐；

【导师】 杨兆选； 王江；

【作者基本信息】 天津大学 ， 信号与信息处理， 2007， 博士

【摘要】 随着科学技术的发展,对非线性伺服系统的控制性能要求也随之提高,传统的线性控制理论已经不能满足要求,近年来,各国学者和研究人员纷纷致力于非线性控制理论和应用的研究,并已经取得了卓有成效的成果。对于非线性系统来说,除极少数情况外,目前还没有一套可行的通用方法,而且每种方法只能针对某一类问题有效,不能普遍适用。因此寻求结合多种控制方法优势的非线性系统综合控制的有效方法成为本论文的主要研究课题。总的来说,伺服系统的控制目的是外加控制使得系统的输出能够较好的跟踪参考输入,且尽可能获得高质量的动态和稳态性能,并且具有较强的鲁棒性能。首先,论文对传统的Backstepping控制进行改进,削弱其应用的强制性假设条件,得到广义的Backstepping控制,在此基础上结合原始模型和规范化模型引入一种易于实现的神经网络自适应在线补偿技术解决系统的鲁棒性问题,并在规范化模型的基础上引入平滑轨迹跟踪滤波器前馈补偿使系统获得高质量的动态性能。其次,针对MIMO系统的Lyapunov渐进稳定性问题,考虑一般形式的参考输入信号,即稳态值为某个稳定的自治系统产生,结合内模理论构建条件伺服补偿器和滑模稳定控制器,在此基础上针对实际系统的未建模非线性部分,引入易于实现的神经网络自适应在线补偿技术较好的解决系统的鲁棒性问题,并引入平滑轨迹跟踪滤波器前馈补偿使系统获得高质量的动态性能。最后,以永磁同步电动机为实际的应用对象,将本论文所提出的两种反馈-前馈复合控制方法应用于永磁同步电动机的位置伺服控制,通过仿真和实验验证所提方法的正确性和有效性。更多还原

【Abstract】 With the in-depth improvement of technology, the conventional linear control theory can not meet the request of higher satisfied dynamic performance of nonlinear servo control systems. Recently, many researchers applied themselves to nonlinear control theory and its applications and acquired many significant results. However, for nonlinear systems, despite of several special occasions, there is no a kind of feasible general method and each control method is only applicable to solve some class of problem. Thus, the main research of this paper is to seek a valid integrated control technique which takes advantages of multi control methods. In general, the control purpose is to make the output track the reference input preferably with the high dynamic, static and strong robust performance by exerting control on the nonlinear plant.Firstly, by weakening the strict suppose for its application, the conventional backstepping control method is extended to general backstepping control method. Then based on the original model and the normal model of system, the robust problem is solved by introducing a kind of Neural Network adaptive online compensation law which is easily to implement in practice. Moreover, based on the normal model of system, good dynamic performance is acquired by adding the nonlinear smooth trajectory filter (STF) feedforward.Secondly, aiming at the Lyapunov stability problem for MIMO nonlinear systems and considering general reference input i.e. its steady value are generated by some natural stable exosystem, the conditional servocompensator and slide mode stabilizer are designed based on internal model theory. Furthermore, the non-modeling nonlinear parts of system are compensated by introducing Neural Network adaptive online compensator and the dynamic performance of system is greatly improved by adding the nonlinear smooth trajectory filter strategy.Lastly, in this paper, considering the permanent magnet synchronous motor (PMSM) as the controlled plant, the proposed two feedback-feedforward integrated control methods are applied to implement the position servo control for PMSM. And the validity and effectiveness of the nonlinear control strategies are verified through the simulations and the practical DSP experiments.更多还原