Condition Assessment of Power Transformers through DGA Measurements Evaluation Using Adaptive Algorithms and Deep Learning

Abstract

Condition assessment for critical infrastructure is a key factor for the wellbeing of the modern human. Especially for the electricity network, specific components such as oil-immersed power transformers need to be monitored for their operating condition. Classic approaches for the condition assessment of oil-immersed power transformers have been proposed in the past, such as the dissolved gases analysis and their respective concentration measurements for insulating oils. However, these approaches cannot always correctly (and in many cases not at all) classify the problems in power transformers. In the last two decades, novel approaches are implemented so as to address this problem, including artificial intelligence with neural networks being one form of algorithm. This paper focuses on the implementation of an adaptive number of layers and neural networks, aiming to increase the accuracy of the operating condition of oil-immersed power transformers. This paper also compares the use of various activation functions and different transfer functions other than the neural network implemented. The comparison incorporates the accuracy and total structure size of the neural network.

Application of back propagation neural network in complex diagnostics and forecasting loss of life of cellulose paper insulation in oil-immersed transformers

“…In summary, the proposed approach has been employed in this work, and it is a good precise approach for transformer diagnosis. It could be utilized as well for periodic testing, with the findings compared over time to monitor the level of gas creation and gauge the actual condition of the transformer 1 , 3 , 4 , 23 .…”

“…The cellulose paper insulation end-of-life requirements are commonly proposed 22 to have DP values of 150–250; below 150, the cellulose paper is considered to be mechanically weak. The assessment of paper insulation for its DP value requires the extraction of a few sheet strips of paper from the investigated transformers 23 , 24 . However, the process remains challenging.…”

Adaptive Stochastic Conjugate Gradient Optimization for Backpropagation Neural Networks

“…Limitations include limited application domains, sensitivity to hyperparameters, complexity, and computational overhead, and the need for further research to explore the applicability of PVR in different environments, optimize hyperparameter settings, and address potential computational overhead concerns. [17] introduce an adaptive conjugate gradient (CG) algorithm specifically designed for deep neural networks. The algorithm leverages the CG method, known for its effectiveness in solving optimization problems, and adapts it to the training of deep neural networks.…”

Power Transformer Fault Diagnosis Using Neural Network Optimization Techniques

“…The insulation system used in an oil-filled transformer comprises mineral oil as the liquid insulating medium and paper as the solid insulating material [12]. The dependability of power transformers is contingent upon the efficiency and effectiveness of the insulating system to endure several stressors, including thermal, mechanical, and electrical ones [4,9].…”

“…The optimization procedure is executed by adjusting the model's hyperparameters. The available literature on hyperparameter tuning is limited, with Machine Learning (ML) algorithm developers only supplying brief descriptions of their functions [9]. While scientific papers and online tutorials offer some additional information, there is a lack of more extensive studies that investigate significant inquiries, such as the extent to which a model can be enhanced using tuning, the most effective tuning strategy, and the influence of tuning a specific hyperparameter on different datasets.…”