连续热镀锌是现代镀锌生产的主要方法,但其镀层厚度系统具有非线性、多变量、时变大滞后等特点,难以建模与控制。本文在对镀锌生产线原理、镀层厚度影响因素分析的基础上,构建与训练镀层厚度预测神经网络模型,该模型具有预测精度高,增益准确,覆盖全工况的优点,为下一步镀层厚度控制系统的开发奠定了基础。 Continuous hot-dip galvanizing is the main method of modern galvanizing production. However, the coating thickness system has the characteristics of non-linear, multivariable, time-varying and large hysteresis, which makes it difficult to model and control. Based on the analysis of the principle of zinc plating production line and the influencing factors of coating thickness, a neural network model of coating thickness prediction is constructed and trained. The model has the advantages of high prediction accuracy, accurate gain and full coverage. System development laid the foundation.