变负荷下超（超）临界机组过热器壁温预测

对超（超）临界机组过热器管壁温度的影响因素进行了分析，利用电厂现场DCS系统采集到的变负荷条件下的运行数据，与对应时刻管壁的温度实测数据进行了关联比较，确定了预测模型的输入变量。分析结果显示：一级、二级过热器出口汽温、主蒸汽温度、二次风E层风箱开度、有功功率等因素对过热器管壁温度的影响较为显著。采用BP神经网络算法，选取关联结果阈值超过0.70的14种主要因素进行升负荷、稳定负荷和降负荷3种条件下的管壁温度预测，预测结果与实测结果整体趋势保持一致，最大相对误差为1.42%，能够对过热器超温预警起到良好的指导作用。

Prediction of Superheater Tube Wall Temperature in Supercritical/Ultra-Supercritical Boilers for Different Loading

In this paper, the influencing factors of the superheater wall temperature in supercritical/ultra-supercritical boilers are analyzed. By using the real-time operation data acquired from the DCS system in a power plant,the grey relational analysis on the measured temperature of the superheater tubes is conducted to determine the input variables of the prediction model. The results show that the influencing factors, such as the outlet steam temperature of both the primary and secondary superheaters, the main steam temperature,the layer E opening of secondary air throttles and the active power are vital to the tube wall temperature. Then by using the BP neural network algorithm, 14 main influencing factors with the threshold of more than 0.70 are used to predict the tube wall temperature at the scenario of up loading, steady loading and down loading,which concludes that the development trend of the prediction results is consistent with that of the measured results, and the largest relative error is about 1.42%. The prediction results can provide good guideline to avoid overheating.