Bearing Fault Identification with the Application of Kalman Filter

For predictive maintenance, it is very important to be able to make predictions as well as possible to avoid stopping the production lines. Thus, we started taking the data provided by the vibration and temperature sensors and with the help of the Kalman filter, we achieved a much clearer trend of bearing wear in the measured systems by eliminating the noises. The Kalman filter is an estimator, a “state estimator” for systems that can be modeled by the linear quadratic problem. This is defined as a problem in estimating the instantaneous state of a linear dynamic system perturbed by white noise. The Kalman filter is also used to predict the probable future course of dynamic systems where there is little chance they will be regulated, such as the course of a river during a flood or the trajectory of a celestial body. In this paper, approached an analysis of vibration signals using the Kalman filter, and after obtaining the data from the vibration sensor, we applied the filter and obtained a better prediction of the signals.