BP神经网络预测与优化棉织物的染色工艺

Prediction and Optimization of Cotton Dyeing Processes by BP Neural Networks

主要利用BP神经网络预测棉织物染色性能和优化染色工艺.神经网络技术与正交实验相结合,建立染色影响因素与染色深度和色牢度关系的BP神经网络数学模型(5-10-2),以正交实验样本训练网络,利用随机样本检验网络的准确和稳定性.在此基础上,利用网络摸拟出全局的染色工艺,对其进行因素水平的取值频率统计,最后得到较优的工艺参数范围.

In this paper, prediction and optimization of cotton dyeing processes are discussed by using BP neural networks. A BP neural network with a 5-10-2 layer structure is built by using orthogonal experimental and BP neural networks. The network is trained with orthogonal experimental samples and tested with random samples. On this basis and by using the trained network, the overall results are obtained by inputting overall processing parameters. By applying frequency statistics to various levels of factors, the optimal process parameters are produced.