Maintenance And Replacement Strategies Of Electric Power Equipment Based On Condition Monitoring And System Risk Evaluation

The traditional maintenance methods that concentrate on equipment itself butignore impacts of equipment failure on system cannot meet the requirements ofadvanced asset management in smart grids. The fundamental target of maintenance is toreduce system risks and increase reliability and economy of power supply. Themaintenance strategy of equipment should be determined by not only its own conditionbut also the effect of its failure on system risk. At the present, system reliabilityassessment does not use equipment condition monitoring data but only uses the averagefailure rate of equipment, and condition monitoring based maintenance and reliabilitycentered maintenance are implemented separately.To overcome the barrier, the thesis presents a repairable failure rate model basedon condition monitoring data and an ageing failure unavailability model based oncondition monitoring data. The maintenance strategy of combining equipment conditionmonitoring and system risk evaluation is proposed using the two models. With such amaintenance strategy, both equipment condition and system risk can be considered in aunified maintenance process. Various different topics have been investigated includingselection of maintenance candidate set, maintenance scheduling, replacement of ageingequipment, economic analysis in maintenance, overall maintenance procedure andmaintenance strategy under power market environment. Maintenances of different typesof equipment in generation, transmission and substation have been studied, and thecorresponding maintenance strategies and calculation methods have been are proposed.The maintenance scheduling method of generators is studied under the powermarket environment. A multi-objective maintenance optimization model based onbidding is proposed. A constraint mechanism to regulate possible abnormal behaviors ofgenerator companies is presented. The NSGA-II algorithm is utilized to solve thegenerator maintenance multi-objective model. This algorithm can avoid the drawback oftransforming a multi-objective to a single-objective in conventional methods. Aresulting non-inferior solution set is used to build the voting strategy for determining thefinal maintenance scheme, which can guarantee the economy of maintenance solutions,system reliability and fairness among generator companies.The maintenance strategy of overhead lines is studied. A new failure rate modelbased on condition monitoring data is presented to establish the quantified relationship between condition monitoring data and system risk. The quantitative maintenanceeffectiveness can be determined. The Maintenance Reliability Benefit Index and theMaintenance Cost Utility Index are proposed. The maintenance ranking order is createdusing a benefit/cost ratio index. An overhead line whose maintenance can improvesystem reliability more will be automatically maintained before a line whosemaintenance can improve system reliability less. The maintenance budget constraint canbe incorporated into the process of an economic and reliable maintenance scheme.The replacement strategy of ageing transformers is studied. The unavailability oftransformer ageing failure is calculated using the relationship between conditionmonitoring data and insulation age so that the system risk can be evaluated bymonitored measurements. The expected damage cost due to ageing failures and thebenefit due to delaying the replacement of transformer are evaluated respectively todetermine the best replacement timing. The proposed method can control the systemrisk due to transformer ageing failures, which avoids either too late retirement ofequipment at important locations or too early retirement of equipment at unimportantlocations.The maintenance of circuit breakers is studied. An evaluation method of circuitbreaker condition is proposed using the information on mechanical parts and secondarycircuits which can be identified by analyzing the waveforms of trip-and close-coilcurrent, waveforms of contact voltage transition, operation times of different auxiliarycontacts, and trip and close initiation time. A non-parametric kernel density evaluationmethod is used to assess the probability distribution of a breaker condition index. Thenthe probability of a circuit breaker operating in the abnormal condition can be assessed.With the abnormal probability and the system risk of substation, the maintenancestrategy of circuit breakers with economy and reliability is presented.