Research On Condition Assessment And Predictive Maintenance Methods Of Power Transmission And Transformation Equipment

With the development of economic society and information technology,people's daily life and production activities are directly affected by the stable supply of power resources.The safe and stable operation of power transmission and transformation equipment is critical to the reliability of the power system.Therefore,the maintenance of equipment is becoming more and more important in the management of power companies.On the other hand,with the development of smart grids and digital substations,the automation level and management requirements of power equipment are constantly improving,which brings great challenges to the maintenance and management of equipment.Therefore,it is of academic discussion and practical application value to study the methods of fault prediction and maintenance scheduling based on the actual operating condition of equipment.At present,the limitations of the mainstream preventive maintenance system that combines regular periodic maintenance and after-sales maintenance have become more apparent in equipment management,such as the increased maintenance costs and unplanned downtime,and it is difficult to meet the development requirements of smart grid.At the same time,with the development and popularization of advanced information detection and storage technologies,the feature data resources related to the operating condition of power equipment are exploding and stored on information sharing platforms,which implies useful knowledge related to the abnormal fluctuations of indicators,failure occurrences and changes in health status.On this basis,this thesis studies the online assessment methods for the current operating condition of power equipment and the prediction method for the fault state in the maintenance interval,and proposed the predictive maintenance methods for single equipment with multiple faults and multiple equipment with single fault respectively.In this thesis,the current operating condition of power equipment is online measured,and the state of potential fault is effectively predicted in maintenance interval,which provides the basis for the scheduling of the equipment maintenance/repair works.Firstly,based on the associated knowledge implied in the actual data resources,a progressive scoring model is proposed to online quantify the fault states and panoramic operating status of different types of equipment with objective score values.Then,considering the different resistance capabilities of the equipment in different health conditions,a prediction model is constructed by categories to measure the deterioration state of potential fault in maintenance interval.Finally,the maintenance priority is divided based on the current state,the changing states in maintenance interval and hazard of potential fault,to optimize resource allocation and maintenance operations.The predictive maintenance problems of single equipment with multiple faults and multiple equipment with single fault are modeled and solved.Through the composition and feature analysis of data resources,circuit breaker and power transformers are selected as the specific research objects.Combined with the change characteristics of fault and operating condition,this thesis studies the assessment methods of current operating condition of equipment,the prediction method of fault state in maintenance interval and the predictive maintenance methods of equipment.Specifically,the main contents of this thesis are as follows.(1)Based on the real-time monitoring data,through the hazard difference analysis of fault occurrence,a neural network is improved to train the influence degree of different faults on the operating condition of circuit breaker,and the current condition information of circuit breaker is quantified by combining the technology of association rule and scoring method in the operation process.The results of sample test show that the accuracy can reach more than 96%,which verifies the effectiveness of this method.(2)For the power transformer with relatively stable condition that operating in normal and stable environment,combined with the data detected at assessment time and periodic pre inspection test data,a scoring function with time effect parameters is constructed to measure fault state,and the operating condition of equipment is assessed based on association rules anf variable weight formula.In addition,considering the polymorphism and fuzziness of state information of power transformer operating in complex operation environment,combined with the periodic pre inspection test data and continuous real-time feature data,the fuzzy functions and Bayesian network are used to improve the scoring model,and the operating condition of equipment is online measured by the scoring value.The test results show that these two methods are suitable for the condition assessment of power transformers operating in different environments.(3)Aiming at the unity and one-sidedness of the existing failure prediction mode that only considers the service age,the hazards of all faults are classified according to the risk description in the technical regulations.Considering the differences in the resistance capabilities of the equipment operating under different conditions,the prediction model of fault state is improved to predict the deterioration trend of potential fault in maintenance interval under the current condition of the equipment.The effectiveness of the model in the prediction of fault states is verified by sample tests and deviation analysis.(4)For the maintenance management of single equipment in the unit inspection interval,a predictive maintenance strategy is proposed based on the fault interval state to improve the existing preventive maintenance strategy.With the preset inspection period,the research is focused on the maintenance/repair problem of single equipment with multiple faults.With different maintenance requirements,the fault state is divided into general quality degradation and severe system deterioration stages by the hazards.Maintenance operations are scheduled according to the current state and maintenance interval states of potential fault.The case test results show that,the proposed method has a good application effect in reducing maintenance costs and unplanned downtimes.(5)For the maintenance scheduling problem of multiple parallel equipment with single fault,the maintenance priority is divided by fault state,and the maintenance time windows are set up based on fault interval state.Considering the constraints of limited maintenance resources and equipment downtime,with the goal of minimizing the total maintenance cost,a predictive maintenance scheduling model is constructed and algorithmically solved to optimize maintenance time and operation sequence.Case test results show that the maintenance scheme obtained by the proposed model can greatly reduce maintenance costs and improve the stability of equipment operation.The above contents are closely linked to form a systematic decision-making framework for the maintenance of power transmission and transformation equipment.Finally,by summarizing the above research results,some meaningful problems in this field that need further research are pointed out.