| Power transformers are key equipment in power systems.If their operating conditions are unstable,they will have a serious impact on the safety and reliability of power systems.Therefore,it is important to evaluate the health status of power transformers to ensure the safety of power transmission and transmission,and to improve the reliability of power systems and promote the development of Condition Based Maintenance(CBM).This paper analyzes the limitations of the Hidden Markov Model(HMM)applied to the transformer field,and introduces the improved Hidden-semi Markov Model(HSMM)into the health state of the power transformer.In the field of evaluation,a method for assessing the health status of power transformer HSMM is proposed,and a power transformer health status detection system is designed and developed.This paper mainly does the following work:1)The research on the mechanism of dissolved gas in power transformer oil is studied,and the degradation mechanism of transformer is studied from the aspects of electrical degradation,thermal degradation,chemical degradation and mechanical degradation.It is pointed out that the current cost of transformer maintenance and repair is too high.It is necessary to carry out the assessment of the health status of the transformer.2)Through the description of the parameters and structure of the HMM model,the feasibility of applying HMM to the field of transformer health assessment is discussed and verified by experiments.The results show that HMM can effectively divide the health of the transformer.3)For the limitation that HMM is insensitive to historical state in the application,introduce the "micro state-macro state" correspondence to construct HSMM,and make multiple observations of the transformer into a "macro state",forward-backward algorithm The redefinition is carried out to incorporate the historical state information of the transformer into the modeling process,and the health and state of the transformer are evaluated by the mean and variance of the residence time of each health state of the output transformer.And verified by experiments,the results show that HSMM improves the correct rate of state division.4)Designed and developed the HSMM-based power transformer health status detection system based on DGA data.The system uses COM technology to encapsulate the HSMM model in MATLAB,and uses the Visual C# development platform to call the packaged HSMM.The workflow,structural framework and main functions of the system. |
