总结了数据挖掘技术在燃煤锅炉故障诊断、燃烧优化、污染物减排及机组优化运行等方面的应用现状，分析了关联规则、聚类分析、神经网络和支持向量机等数据挖掘算法在锅炉优化运行和污染物排放控制中的优缺点。分析表明：人工神经网络鲁棒性强、可自学习且适用面广，未来可基于焚烧机理并耦合其他算法进行工程应用；对于在高控制要求下智能化工况优化空间大的垃圾焚烧锅炉中的发展及应用，建议将数据挖掘技术与云计算平台结合，并考虑垃圾焚烧过程的实际工况和特性进一步开发数据预处理方法，扩大动态数据采集范围，提高模型的实际运行效率和泛化能力。

Based on the advantages and disadvantages of data mining algorithms that applied in coalfired boiler optimization operation and pollutant emission control,including association rules,cluster analysis,neural networks as well as support vector machines,this paper reviews the status on fault diagnosis,boiler combustion optimization,pollutant emission reduction and unit optimization operation by data mining technology. The analysis shows that artificial neural network is relatively widely used due to its strong robustness and selflearning ability. In the future,based on the incineration mechanism and coupled with other algorithms,it can be further applied in engineering. Furthermore,on basis of existing application of data mining technology,the future development and application of the intelligent waste incineration boiler that has largely optimization space under high control requirements have been prospected. We suggest to combine the data mining technology with the cloud computing platform,consider the actual working conditions and characteristics of the waste incineration process,and further develop data preprocessing methods to expand the dynamic data sampling range and improve the actual operation efficiency and generalization ability of the model.

TK229.6，TK11+4

国家重点研发计划项目(2020YFC1910100)