Diagnosis Bearing Faults Based on the Triplet Optimized Embedding Models

Diagnosis Bearing Faults Based on the Triplet Optimized Embedding Models

In rotating machinery, the bearing is one of the important components which improves the rotating machinery’s performance. The bearing quality determines the machine’s performance and reliability. Therefore, fault detection is a key technology to ensure the bearing’s safety and reliability. In the bearing fault diagnosis, separating the sensitive signal from vibration data is one of the challenging tasks due to the large volume of the rolling bearings. The research challenges are overcome using the Triplet Optimized Embedding Model (TOEM) that classifies the faults bearings with maximum accuracy. The triple embeddings are initially created using the ant-optimized long short-term neural model that minimizes the vibration signal. This process extracts the features from the collected data and has been classified using the autoencoder neural model. Encoder, decoder, and activation functions are incorporated during the classification process to classify the faults in bearings. The training process maximizes the fault detection accuracy compared to the existing machine learning classifiers.