基于最小二乘支持向量机的煤粉着火温度预测分析  

Prediction and Analysis of Pulverized Coal Ignition Temperature Based on a Least Square Supportive Vector Machine

作　　者：常爱英[1] 吴铁军[1] 赵虹[1] 包鑫[1] 

机构地区：[1]浙江大学工业控制技术国家重点实验室,浙江杭州310027

出　　处：《热能动力工程》2011年第1期97-99,126-127,共3页Journal of Engineering for Thermal Energy and Power

摘　　要：针对关系到锅炉经济安全运行的煤着火温度估计难的问题,采用最小二乘支持向量机方法建立煤粉着火温度的预测模型,并和利用PLS以及BP神经网络等方法建立的预测模型进行对比,结果表明,最小二乘支持向量机克服了BP神经网络泛化能力弱以及PLS无法解决的非线性等问题,采用最小二乘支持向量机方法建立的煤粉着火温度模型具有很高的预测精度。In the light of the problem relating to the economic and safe operation of a boiler that it is difficult to predict the ignition temperature of coal,the least square supportive vector machine method was used to establish a model for predicting the ignition temperature of the coal and compare it with the prediction models established by using the PLS(partial least square) and BP(back propagation) neural network method etc.The research results show that the least square supportive vector machine can overcome the problems such as weak generalization ability of the BP neural network and nonlinearity that the PLS method has no way to solve.The model under discussion enjoys a very high prediction precision.

关 键 词：煤粉 着火温度 预测模型 最小二乘支持向量机 BP神经网络 PLS 

分 类 号：TQ533[化学工程—煤化学工程]