为了及时发现并处理高压加热器运行经济性失常，采用传热系数直观地反映高压加热器的运行效率，提出基于时序数据分析方法得到传热系数的在线动态模型。首先通过热动力学机理分析得到影响高压加热器传热系数的主要特征参数并建立基于特征参数的动态模型；其次，通过蜻蜓算法改进的密度聚类方法构建具有最优邻域参数的优化聚类模型，得到可信端差区间。通过一段时间的某电厂的计算结果比较表明，基于改进密度聚类法的传热系数在线动态模型在计算高压加热器传热系数时均方误差MSE低至0.030 5％，说明该模型有效、可行。

In order to detect and deal with the operating economic abnormality of high pressure heater in time, this study used heat transfer coefficient to directly reflect the operating efficiency of high pressure heater, and put forward an online dynamic model of heat transfer coefficient based on time series data analysis method. First, the main characteristic parameters affecting the heat transfer coefficient of the highpressure heater were obtained through thermodynamic mechanical analysis, and a dynamic model based on the characteristic parameters was established; second, an optimal clustering model with optimal neighborhood parameters was constructed by the densitybased spatial clustering of applications with noise (DBSCAN) method improved by the dragonfly algorithm (DA) to obtain the credible enddifference interval. Through a period of comparative calculation results of a certain power plant, it is shown that the mean square error (MSE) of the heat transfer coefficient of the highpressure heater is as low as 0.030 5% based on the online dynamic model of the heat transfer coefficient of the DADBSCAN, indicating that the model is effective and feasible.

TH86

国家自然科学基金青年科学基金(51906133)