基于混沌预测的模糊神经网络控制器设计及应用

由于混沌时间序列内部确定的规律性 ,其重构出混沌吸引子的相空间具有高精度短期预测性。根据非线性、大时滞系统的时间序列重构相空间 ,计算相空间的最大Lyapunov指数、饱和嵌入维数和可预报尺度 ,并以此为指导 ,对系统作高精度预测。在此基础上 ,又设计了遗传算法优化的模糊神经网络预测控制器 ,实现了对非线性、大时滞系统高精度的自适应控制。将该控制应用在锅炉过热汽温控制中 ,仿真表明该控制的有效性、准确性和鲁棒性。

Design and Application of Fuzzy Neural Network Controller Based on Chaotic Forecast

Because of the intrinsic reqularits of the chaotic time series, theed phase space of the reconstruct chaotic attractors features high precision short term forecast. Therefore, in order to realize the adaptive control of the nonlinear with big time systemlag, first, the phase space of the chaotic attractors is reconstructed, system's embed dimension, the maximal Lyapunov exponent and forecast measure are calculated by the time series of the nonlinear system with big time lay in this paper, lastly, high precision short term forecast is performed for the system. On this basis, a fuzzy neural network forecast controller optimized by genetic algorithm is designed, thus, high precision adaptive control on the nonlinear system with big time lay is realized. The controller is applied to the overheat steam temperature system, the simulate result proves the controller's validity, veracity and robustness.