Research On Early Warning Method Of Induced Draft Fan Failure Based On Multivariate State Estimate

According to data from the National Bureau of Statistics,the main mode of electricity production in my country is still thermal power generation,and reducing the frequency and number of failures in various links of thermal power generation production has become a hot spot and focus of social research.In the process of thermal power generation,induced draft fan is prone to failure,and has an important impact on the process of thermal power generation,so it is imperative to strengthen the condition monitoring and fault early warning of induced draft fan.In this paper,multivariate state estimation technology is used to monitor the condition of induced draft fan equipment.Through different statistical residual processing methods,the purpo se of abnormal operation state early warning is realized,and the feasibility of sequential probability ratio test method is verified.Firstly,the mechanical structure and common faults of induced draft fan are studied.In order to understand the working module and principle of DCS system and SIS system in thermal power plant,a real-time / historical data acquisition tool is developed based on python,which realizes fast and stable data acquisition in thermal power plant.Combined with the current theoret ical research and field application of multivariate state estimation algorithm,according to the principle of measuring point selection,the process of parameter variable determination based on multivariate state estimation model is explained in detail.Based on Fangjiazhuang power plant,this paper uses multivariate state estimation technology to establish induced draft fan fault early warning model based on multivariate state estimation.In this model,the similarity function is used to realize the function of fault early warning.This paper analyzes the fault early warning principle of the similarity function method,clarifies the model residual processing,and introduces the sequential probability ratio test method to replace the similarity function method to realize the fault early warning function.Through the model verification and field application,The sequential probability ratio test method can better realize the fault warning function of thermal power plant equipment,and supplements the new field application practice for the fault warning method of induced draft fan based on multi-state estimation.