压电陶瓷执行器在高频状态下的辨识与控制

With the development of precision manufacturing technology, the piezoceramic actuators (PEA) often perform trajectory tracking and fast positioning operation in the high-frequency operating condition. However, in this operating condition, the hysteresis characteristics existing in PEA will exhibit more complex properties. Moreover, the dynamic linearity surrounded with the hysteresis nonlinearity also cannot be ignored, which is called as the sandwich system with hysteresis. In this case, both its input and output signals cannot be measured directly because those signals are surrounded by others subsystems. Thus, it is very difficult to use the traditional scheme to identify and control such a system. On the other hand, under the high-frequency operating condition, actuator failure often exists in PEA, i.e. the coupling relationship between the output displacement and output force of the PEA cannot be ignored. In this research, the identification and control of PEA in the high-frequency operating condition is studied. First, an identification scheme is proposed to truly identify each subsystems of the sandwich model with hysteresis. Then, the corresponding sandwich control scheme is designed to perform trajectory tracking and fast positioning operation of the PEA in the high-frequency operating condition. Second, in order to eliminate the actuator failure of PEA, the new control method, which is named as the combined optimized control scheme with the independent constraints, is proposed. In this scheme, a neural network is firstly constructed to identify the coupling model between the output displacement and output force of the PEA. Then, the combined optimized control scheme is designed and applied for the trajectory tracking and fast positioning operation in the high-frequency operating condition, where the identified neural network is served as the independent constraint.

随着精密制造技术的发展，压电陶瓷执行器（PEA）往往需要在高频状态下进行轨迹跟踪和快速定位操作。但是，PEA所含有的迟滞特性在高频状态下表现出更为复杂的特性；而且，迟滞两端的线性动态特性也无法忽视，构成所谓的“三明治系统”。此时，迟滞的输入输出信号成为了中间信号，难以直接测量，这将导致辨识和控制工作更具挑战性。另一方面，在高频状态下，PEA更容易产生“执行器失效”问题，即PEA的输出位移和输出力之间的耦合关系也无法忽略。本项目研究高频状态下的压电陶瓷执行器的辨识与控制问题，提出一种能够真实辨识含有迟滞的三明治系统的建模方法，并设计对应的三明治控制器；为了解决“执行器失效”问题，本研究通过对PEA的输出位移与输出力进行复合优化控制的方式来解决，即：首先利用神经网络模型建立输出位移与输出力之间的耦合模型，并根据压电陶瓷执行器的实际工作特点，采用“带约束优化”的方式来设计对应的控制器。

随着精密制造技术的发展，压电陶瓷执行器（PEA）往往需要在高频状态下进行轨迹跟踪和快速定位操作。但是，PEA所含有的迟滞特性在高频状态下表现出更为复杂的特性；而且，迟滞两端的线性动态特性也无法忽视，构成所谓的“三明治系统”。此时，迟滞的输入输出信号成为了中间信号，难以直接测量，这将导致辨识和控制工作更具挑战性。另一方面，在高频状态下，PEA更容易产生“执行器失效”问题，即PEA的输出位移和输出力之间的耦合关系也无法忽略。本项目研究高频状态下的压电陶瓷执行器的辨识与控制问题，提出新的迟滞模型，并以此为基础提出一种能够真实辨识含有迟滞的三明治系统的建模方法，并设计对应的、基于直接补偿或抵消的三明治控制器；为了解决“执行器失效”问题，本研究通过对PEA的输出位移与输出力进行基于目标交替约束的复合控制策略的方式来解决。最后，本项目还运用机械学、材料科学的相关理论，对长行程压电陶瓷新材料的材料设计与制备学科交叉问题进行研究；而且，基于“精密滚珠丝杆副+力矩电机”宏运动机构，本项目研究了三明治死区系统和三明治间隙系统的辨识与控制问题。本项目的研究成果可以有效提高压电陶瓷执行器在高频工作状态下的运动控制效果，以及有效的解决相应的“执行器失效问题”，为精密制造和金属增材制造等领域提供性能优良的运动执行器。