Condition Assessment And Fault Diagnosis Approaches For Power Transformers

Transformer is the core equipment of energy transmission and conversion in the power system,and operation status of transformers closely related to safety and reliability operation of the power system.Therefore,the further comprehensive study for transformers,to effectively guide the operation maintenance and state overhaul of transformers and significantly reduce the incidence probability of fault,has a great theoretical significance and practical significance.Insulation condition assessment and fault diagnosis of power transformer are studied in this article.The main work are as follows.The existing condition assessment methods of power transformer mostly focus on assessment model and theory research,while neglect temporal and spatial correlation analysis of evaluation indexes.In view of the problem,an insulation condition assessment method of power transformer based on spatio-temporal information fusion was proposed.In view of randomness and fuzziness of data information in the assessment method of power transformer,a multilateral transformer condition assessment system of power transformer is established Variable weight coefficients of factors layer of evaluation model were determined introducing association rules and variable weight synthesizing theory.A fuzzy membership function was formulated to describe the factor layer,and its original basic probability assignment was calculated.The basic probability assignment of factors and cycles in evaluation model was fused based on D-S evidence theory,firstly,basic probability assignment of factors obtained by the assessment factors layer and its weight coefficient were fused based on time domain information fusion,and then the time domain fusion information was used for airspace fusion,finally insulation condition of power transformers was judged combining reliability criterion The results of practical example prove that the proposed method can evaluate the insulation condition of power transformers accurately and effectivelyIn order to solve the problems such as availability of data extraction,better local optimum,gradient to dissipate more efficiently,this paper presents a new method of power transformer fault diagnosis based on stacked auto-encoders and Softmax classifier.The model of power transformer fault diagnosis is established based on stacked auto-encoders and Softmax regression,then each restricted boltzmann machine of fault diagnosis model is optimized by the unsupervised raining with massive unlabeled samples based on k-step contrastive divergence and adjusts parameters of fault diagnosis model by the supervised algorithm.Finally,the power transformer fault types are determined by Softmax regression.Test results show that the proposed method has the very good stability and higher diagnosis accuracy.