【作者】 周希胜；

【导师】 姜培刚；

【作者基本信息】 青岛理工大学 ， 机械设计及理论， 2008， 硕士

【摘要】 高速高精度运动控制是现代机器制造工业的重要研究领域之一,是高速加工中心的核心,对提高生产效率和产品质量具有十分重要的作用。数控系统越来越广泛应用于工业设备中,现已成为现代化工业设备的核心部分和关键技术,其稳定性、快速性、准确性直接影响着整套工业设备的性能指标。数控系统是一复杂的机电一体化系统,存在的扰动、非线性、模型和参数不确定性直接影响着数控系统的稳定性、快速性、准确性。因此,要想进一步提高数控系统的性能指标,来满足现代化工业设备对数控系统提出的更高要求,就必须考虑系统的扰动、非线性和参数不确定性。因此,本课题的研究具有非常重大的理论和现实意义。本文以数控伺服进给系统为研究对象,按照误差避免和误差补偿进行综合研究,确定了进一步提高系统性能的控制策略——基于扰动观测器的伺服控制算法及相对于时间延迟的FNNG伺服控制算法在数控进给伺服系统中的应用。研究了扰动观测器,前馈控制器和反馈控制器的设计方法。提出了一种抗干扰的状态反馈控制器设计方法。对于FNNG混合预测控制器设计,明确了提高预测模型的精度是首要的问题;提出了加权建模新算法及一种综合预测模型,这两种模型和一般灰色模型相比较精度有很大提高,并且将智能信息处理的思路和灰色建模结合起来了;在灰色系统动态分析方面,对含有灰色预测控制器的时间延迟系统的动态行为进行了分析并得到了几个有理论和使用价值的结论;最后提出了时延系统的Fuzzy-Neural Network-Gray混合预测控制。根据伺服进给系统在实际工作过程中的实际情况,利用MATLAB软件和X-Y两维数控进给实验台对所设计综合Fuzzy-Neural Network-Gray的预测控制器效果和抗干扰性能进行仿真和实验验证,结果表明,本文研究的FNNG混合预测控制器对非线性伺服进给系统能够进一步提高数控系统的稳、快、准性能,且有较强的鲁棒性和抗干扰性能。更多还原

【Abstract】 High-speed high accuracy motion control is recognized as one of the most important areas in manufacturing, which is the kernel technique in high-speed machining center laying an important role in increasing the productivity and quality of manufacturing. The numerical control (NC) system is increasingly widely used in the industrial equipment, and it is the core and key technology of the modern industrial equipment. The stability、rapid and accuracy of itself have a direct impact on the performance of the whole industrial equipment. NC system is a complex mechanical and electronic integration system, its disturbance, nonlinearities and unmodeled dynamics have the direct influence on its performance. Therefore, in order to further enhance the NC system performance to meet the higher demands of the modern industrial equipment, we must consider the disturbance, nonlinearities and unmodeled dynamics. So the study has the very great theoretical and practical significance.The thesis takes the NC servo feeding system as object of analysis and study. According to the idea of synthesizing the error avoidance and error compensation to research, the principle of designing disturbance observer is described and its stability is analyzed .An anti-Disturb state feedback controller is proposed to solve the problem in which most of the existing anti-windup schemes. The main problem is the improvement of the precision of grey model . A new weighted modeling algorithm is presented and the weight value learning algorithm is solved, another integrated modeling method is also presented. Two models can improve the precision compared to the traditional algorithm. By using the method of dynamic analysis of origin neighboring region, the condition and the number of critical steps while no oscillating occurred in time delay systems are obtained. The two important conclusions can offer significant theoretical basis for the parameter design in gray predicative controller. A Fuzzy-Neural Network-Gray predictive control algorithm for large time delay system and an integrated predictive control algorithm for the system with ullknown delay time are presented.According to the NC servo feeding system in the working of the actual situation, in order to prove the control effect and the resistance to interference of the Fuzzy-Neural Network-Gray controller, this thesis carries on the simulation and the experiment by the MATLAB and the X-Y NC feeding bench. The results showed, the Fuzzy-Neural Network-Gray controller can effectively decouple the strong disturbance, nonlinearities and dynamic unmodeled for servo feeding system; it can further enhance the steadily, rapid, accuracy performance of the NC system and has the strong robustness and resistance to interference.更多还原