基于极限学习机的燃煤热值预测研究

Study on calorific value prediction of coal combustion based on extreme learning machine

燃煤热值是评价燃煤品质的重要指标之一,快速准确地预测燃煤热值对燃煤锅炉的燃烧优化以及经济运行至关重要。本文提出一种基于极限学习机(ELM)的燃煤热值预测方法,选取煤的水分、灰分、挥发分和固定碳4种工业分析成分作为模型的输入,以煤质高位发热量作为模型输出,建立基于ELM的燃煤热值预测模型,并对107种不同煤进行预测分析。结果表明:ELM模型具有良好的预测能力,模型的拟合度R2在0.98以上,煤质高位发热量预测值的均方根误差为0.29 MJ/kg;与经典线性模型以及BP神经网络模型相比,ELM模型能更准确地预测燃煤热值,且其运算速度快,具有很好的现场应用价值。

Calorific value of coal is one of the important indexes to evaluate the quality of coal. Accurate and rapid prediction of calorific value for coal is very important for the optimal combustion and economic operation of coalfired boilers in power plants. In the study, a prediction method of coal calorific value based on the extreme learning machine(ELM) algorithm is proposed. Four proximate analysis components, namely the moisture content, the ash content, the volatile matter content and the fixed carbon content, are selected as input of the model, and the gross calorific value of coal is taken as output of the model. A prediction model of coal calorific value based on the ELM is established, and 107 kinds of coal are analyzed. The prediction results of the model show that, the ELM model has good prediction ability. The goodness of fit of the model(R^2) is above 0.98, and the root mean square error of prediction(RMSEP) of the calorific value is 0.29 MJ/kg. Compared with the typical linear model and the BP neural network model, the ELM model can predict the calorific value of coal more accurately, and its calculation speed is faster, so it has good field application value.