摘要
为了提高估算煤灰熔点的精度,采用支持向量机结合遗传算法对求解灰熔点问题进行了建模.将灰成分作为输入量,煤灰软化温度作为输出量,用试验数据对模型进行了校验,结果表明,支持向量机模型预测的最大相对误差和平均相对误差分别为7.4%和0.678%,较精确地实现了对软化温度的预测.支持向量机可用于小样本问题的学习,计算速度快,提高了实时处理与反应最新运行工况参数的预测能力.
In order to improve the accuracy of the ash fusion temperature prediction, support vector machine and genetic algorithms were combined to model the ash fusion temperature of coal blends. The chemical compositions of the coal ash were employed as inputs and the ash fusion temperature as output. The results show that the maximum and average relative predicting error are 7.4 % and 0. 678 %, respectively. Support vector machine can find its application in the small sample training, its calculating speed is fast which can improve the real-time processing performance.
出处
《煤炭学报》
EI
CAS
CSCD
北大核心
2007年第1期81-84,共4页
Journal of China Coal Society
基金
国家自然科学基金资助项目(60534030
50576081)
关键词
支持向量机
动力配煤
灰熔点
support vector machine
coal blending
ash fusion temperature