摘要
运行能损模式识别是实现锅炉运行经济性诊断的一项基础性内容。该文研究的主要目的在于建立一种基于人工智能理论的锅炉运行模式在线识别方法。文中对锅炉运行模式的表达、运行模式特征向量的构成、模式识别方法及其应用等相关内容进行了研究,并针对锅炉运行能损模式识别的固有特点, 提出了运行模式广义特征向量和冗余特征参数等基本概念。在此基础上, 建立了基于自组织特 征映射网络(SOM)的锅炉运行模式识别模型。该文的研究结果表明:① 所建立的模式识别模型可以对锅炉的主要运行能损进行有效识别; ② 通过构造广义特征向量可以避免运行工况变动对模式识别效果的影响; ③ 冗余特征参数的引入能够明显提高模型对运行模式的识别能力。文中通过电站锅炉受热面污染及系统漏风等能损模式识别的实例证实了上述方法的有效性。
Pattern identification of energy loss in operation is one of the fundamental matters for diagnosisng economy of boiler operation. This paper aims at proposing an on-line identification method for operation pattern of boiler based on artificial intelligence theory. The paper studies the expression of operation pattern of boiler, the construction of eigenvector of operation pattern, the method of pattern identification and its application, etc, and proposes some basic concepts such as the generalized eigenvector of operation pattern and redundant characteristic parameter. Further more, the pattern identification model of boiler operation based on self-organizing map (SOM) networks is established. The result shows that ① this pattern identification model can effectively recognize the main energy loss in boiler operation, ② the impact of work condition variation on the effect of pattern identification can be avoided by constructing generalized eigenvector, ③introducing of redundant characteristic parameter can obviously improve the ability of model to identify operation pattern. The validity of the method mentioned above is proved by the case studies of pattern identification of energy loss about contamination on heating surface and air leak of system in utility boiler.
出处
《中国电机工程学报》
EI
CSCD
北大核心
2003年第6期204-208,共5页
Proceedings of the CSEE
基金
教育部高校骨干教师资助计划项目(GG-470-10188-1042)
