为了从超临界CO2锅炉实验数据中挖掘出更多的参数信息，应用固体导热计算、工质管路分配计算、炉膛辐射传热仿真等研究方法对气冷壁及其内、外部热边界进行求解，基于DO模型、支持向量机提出了一种气冷壁壁温和炉内烟温预测模型，并根据实验数据进行了计算。该算法的数据源为运行监测数据，可以实现炉内烟气温度分布、壁温及热负荷等物理量的在线求解。研究结果表明：气冷壁导热计算获得的管壁特定点温度与工质温度温差拟合公式准确性高，相关性在0.99以上；气冷壁壁温与外部热边界求解计算中获得热负荷波动情况一致；烟温预测模型的有效性在传热数据库测试中得到验证，其最大相对误差不超过8.5%。

In order to get more paramerter information from experiment data of supercritical CO2 boiler,the research methods of solid heat conduction calculation,pipe flow distribution calculation and furnace radiation heat transfer simulation were applied to solve the gascooled wall and its internal and external thermal boundaries separately.Then a prediction model of gascooled wall temperature and flue gas temperature in the furnace was proposed based on DO model and support vector machine,and calculated according to the experiment data.The data source in this algorithm came from the operation monitoring data,which could realize the online solution of physical quantities such as the temperature distribution of flue gas in the furnace,wall temperature and thermal load.The research results show that through gascooled wall heat conduction calculation,the fitting formula of the temperature difference between the tube wall at specific point and the working fluid is highly accurate with the correlation above 0.99; the gascooled wall temperature accords with fluctuation of the thermal load curve obtained by solving the external thermal boundary; the validity of flue gas temperature prediction model is verified in the testing of heat transfer database with the maximum relative error of 8.5% at most.

TK122

国家重点研发计划项目（2020YFF0218103）