Yang Liu, Zhao Zhongyang, Li Dongjie
School of Automation, Harbin University of Science and Technology, Harbin 150040, China.
School of Automation, Harbin University of Science and Technology, Harbin 150040, China.
ISA Trans. 2023 Nov;142:515-526. doi: 10.1016/j.isatra.2023.08.022. Epub 2023 Aug 22.
Piezoelectric actuators (PEA) are extensively utilized in high-precision micro-measurement and operation. However, owing to the rate-dependent nature of its hysteresis, its accuracy in certain applications would suffer significantly, and the system would become unstable. To address this issue, a new method for developing a feedback control system that can reduce the rate-dependent impacts of PEA on the positioning system is provided using PEA as the research object. This strategy is based on the fractional order Hammerstein model (FO-Hammerstein). For the fractional order model, a novel fractional order integral sliding mode surface is proposed here that can accurately delineate the dynamic characteristics of PEA. This sliding mode surface is composed of a fractional polynomial and an integral term, which can better minimize static errors and monitor reference signals, and it is built using a fractional neural sliding mode control (BP-FSMC) method. The control technique can be extensively utilized in various systems, such as FO-Hammerstein and those described by the fractional transfer function. The research introduces a neural network and an artificial bee colony algorithm (DeC-ABC) that are used to alter the system's parameters. The study's findings reveal that a system with high resilience can follow the signals from both composite and single input sources. Compared with the fractional order sliding mode control approach on the basis of extended state observer, the fractional order sliding mode control method based on single parameter adaptive law and the proportional integral sliding mode control method on the basis of feedforward compensation, this method has a quicker response time and lower tracking error.
压电致动器（PEA）广泛应用于高精度微测量和操作中。然而，由于其滞后的速率依赖性，其在某些应用中的精度会受到显著影响，并且系统会变得不稳定。为了解决这个问题，以PEA为研究对象，提出了一种开发反馈控制系统的新方法，该系统可以减少PEA对定位系统的速率依赖性影响。该策略基于分数阶哈默斯坦模型（FO-Hammerstein）。对于分数阶模型，本文提出了一种新颖的分数阶积分滑模面，它可以准确地描述PEA的动态特性。该滑模面由一个分数多项式和一个积分项组成，能够更好地最小化静态误差并监测参考信号，并且它是使用分数阶神经滑模控制（BP-FSMC）方法构建的。该控制技术可广泛应用于各种系统，如FO-Hammerstein系统和由分数阶传递函数描述的系统。该研究引入了神经网络和人工蜂群算法（DeC-ABC）来改变系统参数。研究结果表明，具有高弹性的系统能够跟踪复合和单输入源的信号。与基于扩张状态观测器的分数阶滑模控制方法、基于单参数自适应律的分数阶滑模控制方法以及基于前馈补偿的比例积分滑模控制方法相比，该方法具有更快的响应时间和更低的跟踪误差。
