火力发电厂锅炉水冷壁壁温预测耦合模型研究  

Study on Coupled Model of Water Wall Temperature Prediction in Fossil Fuel Power Station Boiler

作　　者：杨传山 YANG Chuanshan(Linyi Hengyuan Heating Group Co.Ltd.,Linyi 276000,China)

机构地区：[1]临沂恒源热力集团有限公司,山东临沂276000

出　　处：《工业加热》2024年第7期50-54,共5页Industrial Heating

摘　　要：锅炉水冷壁壁温不均匀分布会导致部分烟气温度过高,使得烟气中的热能无法充分传递给水冷壁,从而降低了换热效率,且壁温不均匀会使局部过热,引起燃烧不稳定,影响燃烧效果。为此,研究火力发电厂锅炉水冷壁壁温预测耦合模型。分析锅炉水冷壁壁温计算原理,计算锅炉水冷壁壁温,以此联合LSTM构建壁温预测耦合模型,确定构建模型结构与锅炉水冷壁壁温预测表达式。采用多点温度数据融合方式提升历史壁温数据与壁温预测结果的精准度,基于分批估计算术平均方法融合水冷壁多点温度数据,计算壁温的平均值与标准差。将水冷壁多点温度数据融合结果输入壁温预测耦合模型中,计算壁温预测误差补偿,获取最终的壁温预测结果,实现锅炉水冷壁壁温预测。实验结果表明:应用本文模型获得的水冷壁壁温预测结果与实际壁温保持一致,壁温预测平均误差仅为1.291%,表明本文模型水冷壁壁温预测效果好,具有较高的预测精度。Uneven distribution of wall temperature in boiler water-cooled walls can lead to excessive temperature of some flue gas,making it difficult for heat energy in the flue gas to be fully transferred to the water-cooled wall,thereby reducing heat transfer efficiency.Moreover,uneven wall temperature can cause local overheating,leading to unstable combustion and affecting combustion efficiency.Therefore,a coupled model for predicting the water wall temperature of boilers in thermal power plants is studied.Analyze the calculation principle of boiler watercooled wall temperature,calculate the boiler water-cooled wall temperature,and use LSTM to construct a coupled model for wall temperature prediction.Determine the structure of the model and the expression for predicting the boiler water-cooled wall temperature.Adopting a multipoint temperature data fusion method to improve the accuracy of historical wall temperature data and wall temperature prediction results,using a batch estimation arithmetic mean method to fuse water cooled wall multi-point temperature data,calculating the average and standard deviation of wall temperature.Input the fusion results of multi-point temperature data on the water-cooled wall into the wall temperature prediction coupling model,calculate the wall temperature prediction error compensation,obtain the final wall temperature prediction result,and achieve boiler water-cooled wall temperature prediction.The experimental results show that the predicted wall temperature of the water-cooled wall obtained by the model in this paper is consistent with the actual wall temperature,and the average error of wall temperature prediction is only 1.291%,it indicates that the model in this article has a good prediction effect on the wall temperature of water-cooled walls and a high prediction accuracy.

关 键 词：锅炉 多点温度数据融合 壁温预测 水冷壁 火力发电厂 耦合模型 

分 类 号：TK311[动力工程及工程热物理—热能工程]