Study On Grading Assessment Model For Power Transformer Operation Condition

As one of key equipment in power system, power transformer's operation reliability is closely related to the safety and stability of power system. For a long time, the main maintenance for power transformer is periodic maintenance, which is regardless of power transformer's condition in service and will lead to"over maintenance"or"owing maintenance". With the development of condition monitoring technology, condition-based maintenance which ground on condition assessment becomes an inexorable trend to replace periodic maintenance. Condition assessment for power transformer is an uncertain problem with multiple attributes. In this paper, an evaluation indices system divided into on-line and off-line evaluation, as well as condition assessment under natural and accelerating impairment mode, was built, and interval extension theory was introduced to study the method of transformer condition assessment.On the base of preventive experiment, including maintenance history, operation records, working surroundings and accessories section also, a layered evaluation indices system was built. It is divided into on-line and off-line evaluation basing on the data collection needs transformer out of service or not. The weight of evaluation index is described as weight A and weight B. weight A is relied on the condition information content and influence on security, and weight B on the reliability degree of measurement and sensitivity degree of reflecting state. Then, it used Analytic hierarchy process (AHP) and Grey correlation technique for order preference by similarity to ideal solution (GC-TOPSIS) to determine the constant weighs of indices. With AHP method the qualitative description for indices relative importance is compute to get quantitative weights, and with GC-TOPSIS multiple weight programmes are synthesized relied on the grey relative degree to ideal and non-ideal programme. At last, it introduced the variable weight model based on balanced function, which modified the constant weights to get the variable weights, according to the deviation of actual datum and normal state.Considering the impairing speed is different between normal and abnormal conditions, the mode of natural and accelerating impairment were purposed. Then it presented the conception of threatening factor, which including threat from dissolved gas analysis (DGA), water content, inner hot temperature and load manner. Then, the problem of choice and synthesis of natural and accelerating impairment in condition assessment was studied. Considering the error from measurement and disturbance, the index data was extended to an interval. On account of incompatibility indices are apt to false assessment in multi-indices assessment, basing on interval extension theory, a matter-element model for transformer condition assessment with multiple attribute was built.The examples show that, the condition assessment model for transformer can synthesize multiple attributes and get a reasonable evaluation. The grading evaluation saves the time out of service and evaluation work effectively, and the problem of condition assessment for transformer in different condition was well solved by synthesis of natural and accelerating impairment mode.