摘要
煤质的元素成分是锅炉热效率计算中必须的原始数据,电厂一般很难自行测定。利用BP神经网络优良的非线性逼近能力,建立较易获得的煤质工业分析数据至元素分析数据的转换模型,并结合遗传算法全局寻优的特点优化BP神经网络的参数,通过算法比较和实例结果分析,表明该模型可以有效、可靠地运用于煤质元素分析的软测量中,并可方便地应用于其他方面。
Element composition of coal which is essential to thermal efficiency calculation, is difficult to self-determination in power plant. This article makes the best use of BP neural network＇s excellent non-linear approximation ability to build the conversion model from proximate analysis data to ultimate analysis data, and genetic algorithm with the characteristics of global optimization is combined to optimize the parameters of BP neural network. Comparison with the pure BP neural network and practical simulation results shows that the model can effectively and re- liably be applied to the soft measurement in the coal ultimate analysis, and can be easily applied to other areas.
出处
《电力科学与工程》
2011年第6期42-46,共5页
Electric Power Science and Engineering
