The Machine Learning Model with Hybrid Pooling Approach For Transmission Line Insulator Classification and Faults Detection

Insulations are essential in order to ensure that transmission lines function normally. Since the dependability of the system is impacted by the unity of the insulating system, routine inspections must be carried out. A numerous researchers develop various systems employing machine learning to identify insulator faults. The developed model used to classify insulators and identify defects in insulators, however, has a lot of space for improvement. Hence innovative CNN model with hybrid pooling techniques has been proposed to classify and detect the various faults in insulators. On the basis of different performance parameters, the proposed CNN model for insulator fault detection has been evaluated at different epochs. The training accuracy of 96.96% & validation accuracy of 95.41% has been obtained at epoch 20 whereas training loss of 28.28% & validation loss of 12.61% has been obtained at epoch 20. It is observed that the proposed model has better results as compared to different state of arts algorithms. The model successfully classifies insulators into four different classes Good Insulator, Missing Cap Insulator, Flash Insulator, Broken Insulator and effectively detects particular class of insulator with a more than 95% of confidence score.