基于LS-SVM的热处理炉钢板温度预报模型

 根据中厚板热处理炉钢板温度控制的工艺特点,建立了一种基于混合核函数最小二乘支持向量机(LS-SVM)的热处理炉钢板温度预报模型,并给出相应的建模步骤。通过LS-SVM模型拟合输入与输出之间的复杂非线性函数关系,以现场生产工艺数据为训练样本对模型进行学习,再选取测试数据样本对模型进行仿真检验。将模型应用于计算热处理炉钢板温度的数学模型中,仿真结果显示,所建立的模型简单,预报能力强,具有广泛的应用前景。

Predictive Model of Steel Plate Temperature in Heat Treatment Furnace Based on LS-SVM

According to the process characteristics of the medium thickness steel plate temperature control in the heat treatment furnace, a new steel temperature predictive model of heat treatment furnace based on the Least Squares Support Vector Machine (LS-SVM) of mixtures of kernels is established; the model also gives considerations to the some problems, such as the existence of local minima and the choice of the number of hidden units in traditional neural network. Based on the new model, the design steps are given. The LS-SVM model is able to fit the complex nonlinear functional relation between the input and output. Roll speed, steel plate length, width, thickness, upper and lower temperature of furnace, surface temperature of steel plate in previous moment are selected as inputs, surface temperature of steel plate in this moment is chosen as outputs. The very good forecast effect is obtained when on-site production process data are taken as the training samples to train the model, and then the data samples of model test are selected to simulate it. The model is used to calculate the steel plate surface temperature of the heat treatment furnace. The result indicates that the model is quite simple, and has the strong predictive ability and a broad prospect of applications.