Current Transformer Status Online Monitoring Platform Based on Big Data

Transformer failures are currently a major issue due to the widespread use of electronic transformers in smart grid monitoring systems. To stimulate regular upkeep practices, and raise grid reliability and efficiency, this work developed a sophisticated fault diagnosis approach for electronic current transformers (ECTs) utilizing analytics from big data. For efficient defect diagnosis in the ECT, the dynamic eagle perching optimized long-short term memory (DEPO-LSTM) technique is proposed. Fault sample datasets for ECTs are collected from big data to train the proposed approach. Min-max normalization is used in data preprocessing to remove noisy or redundant information. From the normalized data, important features are extracted using principal component analysis (PCA). Next, we apply the proposed technique in the framework for fault diagnosis, and the DEPO technique is used to improve the parameters of the LSTM. The simulations are carried out using the Matlab platform to assess the proposed DEPO-LSTM technique for ECT defect diagnostics. Based on the results, we deduced that our proposed approach outperformed other approaches presently in use for diagnosing ECT faults. The DEPO-LSTM algorithm is evaluated in terms of precision (97.15%), recall (97.42%), accuracy (96.23%), and F1-score (96.36%).