Gaussian Process Multi-Class Classification for Transformer Fault Diagnosis Using Dissolved Gas Analysis

Abstract

Dissolved gas analysis (DGA) is widely used for oil-immersed power transformers as a conventional fault diagnosis tool. However, interpretation criteria from DGA assessment often depends on empirical discrimination from a specialist, which can render unreliable or ambiguous diagnoses. Intelligent fault classification algorithms can be implemented to conquer uncertainty in conventional methods, and which require feature learning of transformer condition information data rather than expert experience. In this paper, a Gaussian process multi-classification (GPMC) method is proposed, which uses multiclass recognition with a Gaussian process (GP) and renders an output with a probabilistic interpretation rather than a deterministic guess. The method is investigated using largescale DGA field datasets to improve diagnostic accuracy, and presents reliable incipient fault diagnosis ability. A kernel-based learning algorithm and versatile artificial intelligence (AI) methods, (support vector machine (SVM), artificial neural network (ANN), K-nearest neighbors (KNN), decision tree and logistic regression (LR)) have been used to obtain comparative classification accuracy in comparison to the proposed method: additional comparison is demonstrated between conventional DGA and AI methods. The effectiveness and robustness of the proposed GPMC method are confirmed by experimental accuracy >95%, which illustrates that the proposed method is able to provide superior and reliable diagnoses for operational transformer faults.

Dynamic Incipient Fault Forecasting for Power Transformers Using an LSTM Model

“…The fault rate for the seven future weeks and true health state are tabulated in Table X. To clarify, actual faults were assessed by a hybrid method combining the Roger's ratio and Duval triangle method to construct a fault database, which was proposed in our previous work [18] [31]. In most cases, the transformer condition was healthy; only rare cases with DT faults and some cases with thermal faults, particularly for T3 occurred, which matched the actual fault type distribution.…”

“…A probabilistic neural network was optimized by a bat algorithm to enhance the accuracy of transformer fault diagnosis in [17]. To overcome the shortcomings of black box intelligent models, a probability classification model based on a Gaussian process for transformer multi-class faults was proposed in [18]. Furthermore, a deep learning method based on image processing was presented in [19] for detecting transformer winding faults.…”

“…In addition, the fault classification method is based on fault codes in the relevant IEC standard. In comparison to work in [23], this paper has introduced intelligent classifiers to distinguish between normal state and specific fault classes using large-scale DGA field data [18].…”

“…Although previous work by the paper authors [18] provides an automated tool for power transformer fault classification using large-scale DGA data by intelligent classifiers, it lacked the predictive capability for future failures. To solve this challenge, in consideration of transformer time sequence information, which is often ignored in common fault diagnostic models and traditional DGA interpretation methods, this paper proposes a dynamic fault diagnosis model to predict future health status for timely assessment of incipient conditions, including determination of the insulation system.…”

On the Application of Artificial Neural Network for Classification of Incipient Faults in Dissolved Gas Analysis of Power Transformers

“…Henceforth, it has become particularly significant to identify inchoate transformer faults. Effectual recognition of inchoate fault of oil-submerged transformers can momentously condense the costs coupled with revamping impaired transformers and recovers grid stability and dependability [1][2][3]. The operation of transformers is generally unremitting and is contingent on thermal and electrical strains.…”