液驱并联机构多维力加载系统CMAC-模糊PID控制

CMAC-fuzzy PID control of multi-dimensional force loading system of hydraulic drive parallel mechanism

针对液驱并联机构多维力加载系统非线性、时变和强耦合的系统特性,提出一种改进PID控制策略以提高系统输出精度。通过构建多维力加载系统动力学模型与液压动力元件模型,分析研究耦合力产生机理。将多维力加载时受系统自身结构影响的各通道间耦合力与系统参数摄动等非线性因素视为干扰,采用模糊控制对PID参数进行实时调整,抑制系统非线性干扰以提高跟踪精度;考虑到传统经验法或专家知识库的模糊推理对PID控制参数调整不够精细,提出一种模糊PID与小脑模型神经网络(cerebellar model articulation controller,CMAC)并行的控制策略,利用其自适应能力与快速学习非线性函数能力对模糊PID控制过程进行学习和插补。研究结果表明:与PID控制器相比,CMAC-模糊PID控制器能够有效地解决液驱并联机构多维力加载系统复杂的跟踪问题。

Aiming at the nonlinear, time-varying and strongly coupled characteristics of multi-dimensional force loading system of hydraulic drive parallel mechanism, an advanced PID control strategy was proposed to improve the output accuracy of the system. The mechanism of coupling force generation was analyzed and studied by building dynamic model and hydraulic power component model of multi-dimensional loading system. Nonlinear factors such as coupling force between channels and system parameter perturbation, which were affected by the system’s own structure, were considered as interference. Real-time adjustment of PID parameters was achieved by fuzzy control, in which nonlinear system interference was suppressed and tracking precision was improved.Considering that fuzzy reasoning based on the traditional experience method or expert knowledge is not fine enough to adjust PID control parameters of fuzzy reasoning, a fuzzy control strategy of PID neural network parallel to the cerebellum model was proposed by using the adaptive ability and fast learning ability of nonlinear function of fuzzy PID control process in learning and interpolation. The results show that compared with the PID controller, the CMAC-fuzzy PID controller can effectively solve the complex tracking problem of the multidimensional force loading system of the hydraulic drive parallel mechanism.