A Multinomial DGA Classifier for Incipient Fault Detection in Oil-Impregnated Power Transformers

Abstract

This study investigates the use of machine-learning approaches to interpret Dissolved Gas Analysis (DGA) data to find incipient faults early in oil-impregnated transformers. Transformers are critical pieces of equipment in transmitting and distributing electrical energy. The failure of a single unit disturbs a huge number of consumers and suppresses economic activities in the vicinity. Because of this, it is important that power utility companies accord high priority to condition monitoring of critical assets. The analysis of dissolved gases is a technique popularly used for monitoring the condition of transformers dipped in oil. The interpretation of DGA data is however inconclusive as far as the determination of incipient faults is concerned and depends largely on the expertise of technical personnel. To have a coherent, accurate, and clear interpretation of DGA, this study proposes a novel multinomial classification model christened KosaNet that is based on decision trees. Actual DGA data with 2912 entries was used to compute the performance of KosaNet against other algorithms with multiclass classification ability namely the decision tree, k-NN, Random Forest, Naïve Bayes, and Gradient Boost. Investigative results show that KosaNet demonstrated an improved DGA classification ability particularly when classifying multinomial data.

Improved intelligent methods for power transformer fault diagnosis based on tree ensemble learning and multiple feature vector analysis

“…Based on IEC 60599, the two main types of faults can, according to their severity, be divided into six faults, as summarized in Table 1. DGA is based on the principle that under normal operating conditions, oil-immersed power transformers generate little or no fault-related gas [13]. However, different fault-related gases can be produced depending on the type of fault and its location.…”

“…Although these methods are simple, easy to understand, easy to implement, and widely used by transformer maintenance professionals [12], they have certain drawbacks in terms of precision and uncertainty. In addition, these methods also have the disadvantage of leading to inconclusive assessments of fault severity, or in the extreme case, misclassification [13,14]. To overcome the drawbacks and improve the accuracy and efficiency of traditional methods, intelligent DGA-based methods have been developed.…”

An Efficient LoRa-Enabled Smart Fault Detection and Monitoring Platform for the Power Distribution System Using Self-Powered IoT Devices

“…where W is the total wattage of the solar panels (3.5W) and V is the battery voltage (4.8V). To cater for inefficiencies in the charging system, a higher battery capacity is assumed and calculated using (8). A battery charge efficiency of 85% is taken, and a new value for the battery charge is obtained using (6).…”

“…For best results, this quantity should ideally be of very low value. The receiver sensitivity (Ѕ) is obtained using (8).…”

“…In the event of a power failure, the service provider remains completely oblivious to the fact, resulting in a hue and cry from the affected consumers. This lack of feedback also delays the initiation of reparative and restorative activities and interrupts the revenue stream for the power system operator (PSO) [8], [9]. In response to this issue, service providers have had to employ strategies that attempt to keep them informed about problems on the distribution side of the network.…”