基于小波神经网络的抄纸过程建模仿真研究

抄纸过程中水分定量控制是一种复杂的多变量耦合过程.它具有大滞后、非线性、时变、不确定等特性,因此一般采用常规方法很难准确建立其数学模型.神经网络是一种很好的非线性数学模型,运用多输入多输出的小波神经网络对抄纸过程中水分定量控制进行数学建模,并在小波神经的嘲络训练过程中采用了两种训练方法,加快了训练速度.结果表明,小波神经网络输出值与实际模型输出值的误差在允许范围内,可以很好的逼近抄纸过程的数学模型.

A Simulation Study on Papermaking Process Based on Wavelet Neural Network

Quantification and moisture control in paper process is a complicated multivariable coupling object.With the characteristics of large time-delay,nonlinearity,and time varying,uncertainty etc.it is hard to accurately establish a mathematical model for it using the conventional method.The neural network is a quite good nonlinear mathematical model.In this paper,MIMO wavelet neural network is used,and in the MIMO wavelet neural network processing,we adopt two training methods,to accelerate the speed.The result...