Vibration-Based Condition Monitoring of Shaft Bearing Systems Using Machine Learning Techniques

A shaft-bearing system is an essential part of rotating machinery. To guarantee that a shaft bearing system operates safely and reliably, the bearings' condition must be monitored on a regular basis. Bearing and shaft failures are thought to be the leading reasons of failure in various revolving machines used in the industry at highre and lower speeds. The condition of the bearing changes throughout use, so do the vibrations, and their characteristics vary depending on the reason. As a result, the bearing's unique property makes it suited for vibration monitoring and other procedures. The vibration measurement approach may reliably anticipate the upcoming failure and life of a mechanism or component based on changes in vibration signals.  As a result, the bearing's unique property makes it suited for vibration monitoring and other procedures. The vibration measurement approach may reliably anticipate the future failure and life of a machine or component based on changes in vibration signals. As a result, the goal is to extend the machine's life by detecting faults early on, allowing for an effective maintenance program to be implemented to remedy the problem. Subsequently, this research uses machine learning methods to detect bearing problems, compare them to various faulty and standard models, and categorize the bearing type. In this research work, we use outer race fault data from the Bearing data set to extract the time domain features from the dataset using Various machine learning models, including Principal Component Analysis, K-NEAREST NEIGHBOURS (K-NN), SUPPORT VECTOR MACHINES (SVM), RANDOM FOREST CLASSIFIER, DESICION TREE, and LOGISTIC REGRESSION. As a consequence, we obtain the best model that performs optimally on the data set. Finally, the proposed methods of condition monitoring will be implemented in a real-world case study of the shaft bearing system. Thus, vibration testing is used to monitor the state of the shaft bearing system, allowing for the identification of problematic bearings and improved performance after they are replaced.