基于高维小波神经网络的热连轧板材质量模型研究

热连轧板材的生产要经过多道工序，影响板材质量的因素众多，且与板材质量之间是极其复杂的非线性关系．本文采用LVQ（Learning Vector Quantization，学习矢量量化）神经网络对热连轧板材的生产样本数据聚类，然后按照热连轧板材的生产工序构建基于双输入层高维小波神经网络的热连轧板材质量模型，将板材炼制的输人参数置于网络的第一输入层，将板材轧制参数置于网络的第二输入层．仿真结果表明，改进的模型拟合命中率达到92．3％，测试命中率达到84．5％，可以满足企业生产的需要．

The Quality Model Research of the Hot-rolling Plate Based on High-dimensional Wavelet Neural Network

The production of hot-rolling plate has multi-channel processes and there are many factors that affect the quality of the plate. They have a complex nonlinear relationship between the factors and the quality of the plate. Learning Veeeor Quantization （LVQ） neural network was used to cluster for sample data in the paper. According to the production process of hot-rolling plate, the structure of large-dimensional wavelet neural network with double input layer was used to build the quality model of hot-rolling plate. The refining input parameters were placed in the first input layer of network, and rolling input parameters were placed in the second layer. The simulation results showed that fitting hit rate of the improved model was 92.3% , and testing hit rate was 84.5%, which can meet the need of enterprise produetion.