基于ANSYS与BP神经网络的轧花速度预测方法

The Prediction Method of Saw Gin Speed Based on ANSYS and BP Neural Network

为实现对轧花过程中轧花速度的预测以及在生产中选择合适的速度值,以达到优化原棉品质指标的目的,设计了BP神经网络模型.采用ANSYS软件对轧花过程进行模拟,得到锯片在不同转速下的支反力,并以南疆地区的机采棉为研究对象,以回潮率、支反力以及短绒率为输入量,以轧花速度为输出量,建立了BP神经网络预测模型并对网络进行训练和测试.结果表明:该预测模型可以很好地对轧花速度进行预测,平均预测误差率低于1%.由此表明,可以通过调节轧花速度来提高原棉品质.

In order to realize the prediction and select the appropriate value of saw gin speed to optimize the quality of cotton in the production, BP neural network model was designed. The reaction forces of saw blade under different rotational speeds were achieved through the ginning process simulation in ANSYS. Then taking cotton picked up by machine in South Xinjiang as research object, BP neural network forecast model was built with moisture regain, reaction forces and short fiber content as inputs, and ginning rate as output. The results show that the prediction model can well forecast saw gin speed and the relative average error is below 1%. It is considered that the quality of raw cotton can be improved through the saw gin speed adjustment.