基于多信号源的神经模糊Hammerstein-Wiener模型研究

面对复杂工业过程控制的需求, 设计一种结合数据信息的特殊模型结构, 在保证控制系统有效性的前提下通过模型的结构来简化控制器的求解是亟待解决的问题. 为此, 本文提出一种基于多信号源的神经模糊Hammerstein-Wiener模型, 突破传统的迭代分离方法, 通过组合式多信号实现Hammerstein-Wiener模型中神经模糊非线性环节和线性环节的分离, 同时设计了神经模糊模型参数的非迭代优化算法, 将研究结果拓广到分段非线性系统,改善了模型的适用范围. 该算法保证了模型的预测精度,具有逼近较强非线性过程的能力. 在此基础上设计了基于神经模糊Hammerstein-Wiener模型的控制系统, 利用模型的特殊结构将非线性系统的控制问题简化为线性系统的控制问题, 采用简单的PID控制器便能达到较好的控制效果.仿真结果验证了上述方法的有效性.

Research on Multi-signal Based Neuro-fuzzy Hammerstein-Wiener Model

In order to solve the control problem of complex systems, it is important to design a special structure model with data information to simplify the question of designing control system. Thus, a multi-signal based neuro-fuzzy Hammerstein-Wiener model is proposed, which breaks through the traditional iterative separation method. The separation of the neuro-fuzzy nonlinear and linear parts of the Hammerstein-Wiener model is realized by one kind of multi-signals. And a noniterative neuro-fuzzy optimization algorithm is designed to expand the research results to piecewise nonlinear system, which can be applied to much more nonlinear systems. This algorithm guarantees the precision of the model. Moreover, it has the ability of approximating strong nonlinearity. Furthermore, a neuro-fuzzy Hammerstein-Wiener model based control system is designed to simplify the control problem of the nonlinear system into the problem of linear system by using the special structure of the model. As a result, the traditional PID controller can get a better control result. Simulated results show the effectiveness of the method.