基于核向量机的超超临界机组水冷壁温度预测

Prediction of Water Wall Temperature in Ultra-Supercritical Units Based on Kernel Vector Machine

针对超超临界火电机组水冷壁温度预测问题,提出了基于核向量机预测水冷壁温度模型,并与支持向量机模型进行了对比分析。首先,对超超临界火电机组水冷壁的结构和工作原理进行了介绍。其次,分别介绍了核向量机和支持向量机的基本原理和实现方法,并对两种模型进行了训练和测试。最后,通过实验比较了两种模型的预测精度和效率。试验结果表明:核向量机具有更高的预测精度最大误差为2.02%,其平均误差为0.55%。

This paper addresses the issue of predicting water wall temperature in ultra-supercritical thermal power units by proposing a temperature prediction model based on kernel vector machines,and it is compared with a support vector machine model.Initially,the structure and working principle of the water wall in ultra-supercritical thermal power units are introduced.Subsequently,the basic principles and implementation methods of kernel vector machines and support vector machines are described,and both models are trained and tested.Finally,an experimental comparison is made of the prediction accuracy and efficiency of the two models.The results demonstrate that the kernel vector machine has a higher prediction accuracy with a maximum error of 2.02%,and an average error of 0.55%.