基于LSO-SVR算法的多温区温度偏差预测模型

Prediction model of temperature deviation in multi-temperature zone based on LSO-SVR algorithm

针对多温区温度控制过程中,受到耦合效应及滞后因素的影响,温度偏差难以预测的问题,提出了一种基于狮群优化算法(lion swarm optimization algorithm, LSO)优化支持向量回归(support vector regression, SVR)算法的多温区温度偏差预测模型,利用狮群算法的寻优能力优化支持向量回归的惩罚系数和核函数参数,选取RBF高斯径向基函数作为核函数,建立了多温区温度偏差预测模型,选取多温区实验台的加热棒温度作为预测模型输入,目标温区的温度偏差作为预测模型输出,并将预测结果与粒子群算法优化支持向量回归模型的预测结果进行对比。结果表明,经过狮群算法优化支持向量回归的多温区温度偏差预测模型,在拟合和预测精度上要优于粒子群算法优化的温度偏差预测模型。

Aiming at the problem that temperature deviation is difficult to predict due to the influence of coupling effects and hysteresis factors in the process of temperature control in multi-temperature zones, a method for predicting temperature errors in multi-temperature zone based on the lion swarm optimization algorithm was proposed to optimize the support vector regression algorithm.Using the optimization ability of the lion colony algorithm to optimize the penalty coefficient and kernel function parameters of the support vector regression, the RBF Gaussian radial basis function was selected as the kernel function, and the temperature error prediction model in the multi-temperature region was established.The rod temperature was used as the input of the prediction model, the temperature deviation of the target temperature zone was used as the output of the prediction model, and the prediction results were compared with those of the support vector regression prediction model.The results show that the prediction model of temperature deviation in multi-temperature zone optimized by lion swarm algorithm with support vector regression is better than the prediction model optimized by particle swarm algorithm in terms of fitting and prediction accuracy.