Power Transformer Oil Chromatogram On-line Monitoring And Warning System Research

The normal operation of the power transformer,security and stability of power system plays an important role.In order to guarantee the normal operation of the power transformer,the need for transformer condition monitoring,the operation of the transformer in the process of historical data to predict the change trend,through the change trend,the longitudinal use of transformer on-line monitoring data,based on the dissolved gas in transformer oil change trend to determine the best maintenance time.Transformer oil chromatogram on-line monitoring due to the complexity of maintenance,equipment is expensive,and the smooth operation of the transformer working condition,the variation in content of gases dissolved in transformer oil is not large,the process of gas dissolved in transformer oil with time cumulative effects,monitoring data for a short period of time did not achieve the desired effect of fault prediction and trend from the characteristics of the working condition of the normal operation of the transformer,the phase space reconstruction algorithm,and optimizes the online monitoring cycle,decrease the number of on-line monitoring system for collection,is beneficial to prolong the lifecycle of the online collection system,reduce equipment investment cost.Improved mutual information fuzzy support vector machine(SVM)have good fitting ability to smooth data,through the fuzzy support vector machine(SVM)can be improved mutual information of transformer oil chromatographic stationary under normal working conditions data to estimate that the predictive results of the prediction system combined with early warning system,when the transformer into the non-stationary state,which can realize transformer early warning of abnormal conditions.Transformer accumulated in phase,the hidden danger in an s-shaped curve,the gases dissolved in transformer oil by using the theory of logistic and gompertz combination forecast system can accurately describe the s-shaped growth characteristics,the characteristics of abnormal conditions for transformer oil chromatographic data,and calculate the oil chromatogram data to pay attention to the moment when,optimal maintenance time estimates.The main work of the paper is as follows:(1)in view of transformer oil chromatogram on-line monitoring due to the complexity of maintenance,equipment is expensive,and the smooth operation of the transformer working condition,the variation in content of gases dissolved in transformer oil is not large,the process of gas dissolved in transformer oil with time cumulative effects,monitoring data for a short period of time did not achieve the desired effect of fault prediction and trend from the characteristics of the working condition of the normal operation of the transformer,the phase space reconstruction algorithm,and optimizes the online monitoring cycle,decrease the number of on-line monitoring system for collection,is beneficial to prolong the lifecycle of the online collection system,reduce equipment investment cost.(2)the analysis of the modified mutual information fuzzy support vector machine(SVM)have good fitting ability to smooth data,through the fuzzy support vector machine(SVM)can be improved mutual information of transformer oil chromatographic stationary under normal working conditions data to estimate that the predictive results of the prediction system combined with early warning system,when the transformer into the non-stationary state,which can realize transformer early warning of abnormal conditions.(3)considering the transformer after early warning system for early warning,heralding the transformer into the abnormal condition,will present the S type curve of oil chromatographic data growth model,fuzzy support vector machine(SVM)to predict system improved mutual information will not be able to meet the prediction requirements,to establish logistic and gompertz principle combination forecast system,the system can accurately describes the characteristics of the model S growth characteristics,and realize the abnormal condition of oil chromatographic data,and calculate the oil chromatogram data to pay attention to the moment when,optimal maintenance time estimates.(4)on the basis of the above principle,the design of the transformer on-line monitoring data acquisition system,online monitoring system cycle optimization and adjustment,stable operation condition improved mutual information under the fuzzy support vector machine(SVM)to predict system,early warning system,the logistic smooth operating condition and gompertz principle combination forecast system.Realize online monitoring of transformer oil chromatogram data,monitoring cycle optimization adjustment and fault early warning,the fault trend prediction,and the determination of optimal maintenance time.