关键词: 火力发电机组, 热工参数, 软测量, 支持向量机, 动态, 在线更新

Abstract: In this paper, an algorithm for typical thermal parameters of utility boilers is proposed based on the least square support vector machine(LSSVM), in which the core issues, such as dynamic soft modeling and on-line updating of thermal parameters are intensively explored. Based on the boiler practical operation data, the proposed algorithm is verified with the carbon contents in the flying ash and the oxygen contents in the flue gas taken as cases. A software platform is designed and applied to the fly ash carbon and flue gas oxygen contents measurement for application verification. The software platform is tested in a 600-MW unit in Guangdong. The results prove the validity of the algorithm and the corresponding software platform.

Key words: thermal power generation unit, thermal parameter, soft measurement, least square support vector machine, dynamic state, on-line updating