Simulation Research Of Reactor Fault Characteristics And Comprehensive Evaluation Of Operating Status

Dry-type air-core reactors are important auxiliary equipment in power transmission and transformation systems,and have very good effects in reactive power compensation and short-circuit current limitation.Due to the long-term exposure to external abnormal factors,the internal insulation is deteriorated and damaged,and the reactor fires and burns from time to time.Therefore,accurate and effective status evaluation of power equipment is of great significance to the safety of transmission and transformation systems.This article starts with the analysis of the mechanism characteristics and state influence of common failures of dry-type hollow reactors.Based on the relevant evaluation guidelines and technical standards of reactors,combined with the method of prior simulation calculation and analysis of failure characteristics,objective extraction can be effective State evaluation indicators that characterize the performance of the reactor,and find the correlation between the amount of indicators and the degree of deterioration of the operating state,based on the method of multi-index data information fusion to establish a model to evaluate the operating state of the reactor,and further evaluate the reactor with less than ideal results Through fault diagnosis,the device uses intelligent algorithms to determine the possible types of faults.The main research contents of this article are as follows:(1)Analyze the common fault types and their causes during reactor operation,and summarize the state quantities that can characterize the operating state.Among them,the operating parameters,operating environment and operating history can be characterized from different angles.Status,with reference to the relevant technical standards of reactors and guidelines for evaluation procedures,extract 8 status evaluation indicators in the operating environment and operating history and determine the evaluation area of the indicator data.(2)The average temperature rise and electrical parameters of the reactor during operation are the intuitive reflections that characterize the internal changes of the reactor.The simulation software is used to establish the finite element calculation model of the electrical parameters of the reactor during typical faults,and the fault characterization ability of the electrical parameters is objectively extracted.The output current,active power and power factor are used as key state evaluation indicators to alleviate the problem of strong subjectivity in the selection of state indicators.Further establish the temperature field simulation calculation model at the time of failure,find the change range of the corresponding electrical index quantity when the operating temperature of the reactor changes from normal to overheating,that is,the evaluation domain of the index;find out the law of the average temperature rise of the reactor with the current change,which is the evaluation model The calculation of the weight of the index provides an objective basis.(3)Establish a reactor operating state evaluation model.On the basis of objectively and rationally selecting 12 status indicators in 3 categories,such as operating parameters,operating environment and operating history,and determining the evaluation area of the indicator data,the fuzzy comprehensive evaluation principle combines multiple indicator information to establish an evaluation model.In the weight calculation,combining subjective experience and objective reality,the relationship between the average temperature rise calculated by simulation and the change in current is used to optimize the importance ratio parameter in the index scale to calculate the subjective weight,which is calculated by the correlation method of index importance the objective weights are integrated to make the calculation of the weights more in line with the actual project.After calculating the weight and the degree of membership,the deterioration level of the operating state of the reactor i s obtained.The effectiveness and applicability of the evaluation method in this paper are verified by examples.(4)Adopt artificial bee colony and support vector machine to establish the reactor fault diagnosis model,and further diagnose the possible fault types for the reactor whose evaluation result of the higher-level evaluation model is serious running state.The electrical parameters such as current,active power and power factor that are strongly related to the typical fault of the reactor are used as the input of the diagnostic model,and the overcurrent,the short circuit between different turns and the short circuit between the same turns are used as the model output to determine the fault type.Experiments show that the classification accuracy of the model on the test set data is higher than that of the traditional SVM method,reaching 89.5%.