Research On Degradation Of Motor Based On Multidimensional Operation-and-maintenance Data Fusion

The three-phase asynchronous motor is the most common equipment in power plant,of which the function is to provide motivation to the system by transforming electrical energy to mechanical energy needed in mission accomplishment.Accurate estimation of the motor’s remaining useful life(RUL)is an important prerequisite to ensure the perfect function of the system,reasonable maintenance decisions,and to avoid failure losses.In reality,the data collecting process may encounter some problems like technique difficulties,high costs and so on.However,amounts of data that can be easily collected during the process of operation and maintenance,is related to degradation closely,but rarely used in degradation research.To improve the utilization rate of information and explore a new way to degradation estimation,this article conducts a research on the degradation rules of the three-phase-asynchronous motor based on multidimensional operation-and-maintenance data fusion.The main contents are as follows:(1)Operation-and-maintenance data was analyzed and classified.For the large scale and difficult process problems of initial data,the connotation of the operation-and-maintenance data was defined.The data was categorized according to three aspects including property,source and collecting time.The data was also classified based on ontology,which ensured the intuition and clarity of the data type.(2)The correlation between failure mode and operation-and-maintenance data was analyzed.For the difficulties faced in degradation analysis caused by the variety of failure modes,the information of motor failures was collected.The representative mode was chosen and the corresponding performance was used as the health indicator of the motor degradation.The correlation of data and health indicator was analyzed using stress interference matrix,and then confirmed the types of data needed in degradation prediction.(3)An experiment system was designed.Sufficient data was needed in the research of degradation rules.To solve the problem of data source,an experiment system was designed based on the operation-and-maintenance data obtained by correlation analysis.The operability and economy were considered in the design.The operation,data collection and storage process were determined.Finally,the required operation-and-maintenance data base were produced.(4)Degradation estimation based on data fusion was explored.To avoid the difficult process of sensitive data collection,an addition fusion model and a multiplication fusion model were put forward to describe the degradation process separately.The particle filter was used to train the parameters of the model,and the RUL of the motor was estimated.Finally,The prediction results of traditional method and data fusion method were analyzed comparatively.