Classification of electrical faults in thermal images via deep learning approaches

Asynchronous motors are thoroughly preferred in industrial applications regarding their advantages in comparison with other motor types. Transformers constitute another oft-used and inevitable category for the adjustment of voltage to be fed to an electrical system. The fault diagnostics of these equipment are generally fulfilled by in-depth determinations of stator current signals, magnetic flux distributions, etc. which require comprehensive analyses of conventional electrical measurements. Herein, thermal image analyses arise as an easy way to identify the situations of electrical equipment in which there is no need for direct intervention to the structure. In thesis study, we handle the thermal image-based analyses to distinguish the situations of asynchronous motors and transformers. For this purpose, twenty different situations were obtained combining both asynchronous motor and transformer conditions. These conditions were defined as the cooling fan failure, rotor fault, short-circuit faults in different phases and in various rates inside the motor, short-circuit faults in different rates inside the transformer, no-load motor, and no-load transformer. Without a pre-processing step, classification process was performed utilizing efficient deep learning architectures (DenseNet201, MobileNetV2, ResNet50, ShuffleNet, Xception). In experiments, the hyperparameters of models are examined in a comprehensive manner to observe the highest scores of architectures that can be achieved. Herein, four phenomena (mini-batch size, learning rate, LRDF, optimizer type) are evaluated to perform the transfer learning task, not to spoil the main part of the models, and to reveal the appropriate model for thermal image classification. In trials, an 80%-20% training-test split is allowed to compare the models without data augmentation. As a result, the highest performance is observed with all deep learning architectures by attaining 100% accuracy for condition discrepancy of thermal images belonging to the three phase-motor and one phase-transformer. In addition to the accuracy-based analyses, an in-depth evaluation is presented to reveal the most appropriate architecture in thermal image classification. Regarding this, it's advised that ShuffleNet architecture seems as the most remarkable system among other transfer learning methods.