论文部分内容阅读
通过对热驱动部件的理论分析,确定参数待定的热驱动部件数学模型。针对现有数学模型中多阶指数函数拟合算法的复杂性,提出一种基于函数链神经网络(FLANN)的多阶指数函数逐级递推式拟合算法。根据不同时间常数的指数函数具有不同平衡时间的特点,采用基于FLANN一阶指数函数拟合算法,逐步确定多阶指数函数对应项的待定参数,建立热驱动部件的数学模型。应用多模控制和模糊切换策略,对具有大进给力的纳米级驱动部件进行控制,试验表明系统具有快速响应和高精度,并实现了平稳过渡,证明了基于FLANN的算法构建的控制模型具有精度高、收敛性好以及简单实用等优点。
Through the theoretical analysis of the heat-driven components to determine the parameters to be determined mathematical model of the heat-driven components. Aiming at the complexity of multi-order exponential function fitting algorithm in existing mathematical models, a novel multi-order exponential function recursive fitting algorithm based on functional chain neural network (FLANN) is proposed. According to the characteristic that the exponential functions of different time constants have different equilibrium time, the FLANN first-order exponential function fitting algorithm is adopted to determine the undetermined parameters corresponding to the multi-exponential function gradually and a mathematical model of the heat-driven components is established. The multi-mode control and fuzzy switching strategy are applied to control the nanoscale driving components with large feed force. The experiments show that the system has fast response and high precision and smooth transition. It is proved that the control model based on FLANN algorithm has the accuracy High, good convergence and simple and practical advantages.