About Oilless Bearing Remaining Life

Due to the nonlinear process of bearing degradation, an unscented Kalman filter( UKF)-based approach is proposed to predict the remaining useful life( RUL) of the bearing. The approach includes two parts including bearing performance degradation assessment and RUL prediction.

graphite impregnaed bushings
Oilless Bearing Remaining Life

Bronze Bushing

A key component | About Oilless Bearing Remaining Life

Bearing is widely used in rotating machinery as a key component, but it is also a relatively fragile component in rotating machinery and often fails. Once the bearing fails, its running accuracy will drop sharply, and its performance will inevitably decline, which will cause its main unit (such as aero-engine, CNC machine tool, wind turbine, etc.) to fail to work normally. The health status of the bearing will directly affect the performance of its main unit, so it is necessary to diagnose and predict the fault of the bearing. By judging the running state of the bearing and predicting the remaining life, serious problems such as failure of the main unit, shutdown maintenance, loss of productivity, casualties, and other serious problems caused by bearing failure can be effectively avoided. A reasonable performance inspection plan and maintenance strategy provide a strong guarantee.

Bearing life prediction method

The current life prediction methods for bearings and other equipment can be divided into two categories: data-driven and failure-based physical models. Considering the complex structure of the equipment, the changeable operating status, and the unclear failure mechanism, it is usually difficult to construct a single physical model of equipment failure. a more feasible way.

Oilless Bearing Remaining Life
oil-free bearings

Life cycle of oil-free bearings

The entire life cycle of the bearing can generally be divided into three stages: the running-in period, the effective working period, and the recession period. The running-in period generally refers to the initial operation stage of the bearing. During this period, the various parts of the bearing experience a process of contact, friction, etc., until each part eliminates the deviation during processing and adapts to each other to achieve a stable working state. Therefore, the vibration of the bearing generally changes from strong to weak at this stage. Through the initial running-in, a good balance is achieved between the components of the bearing, the bearing enters a normal working state, and its vibration signal is relatively stable. Due to the harsh working environment, long-term operation, degradation of lubrication performance, etc., the parts where the bearing is more stressed will gradually suffer from fatigue damage, resulting in cracks, shedding, and other failures, and the performance begins to decline until the bearing fails completely. Vibration signal analysis is a more effective method in bearing fault diagnosis and prediction technology. 

Better use of bearing vibration signals

Using the vibration signal of the bearing to construct an index reflecting its health status, clearly indicates that the bearing has experienced three stages of the running-in period, effective working period, and decay period in the whole life cycle. Through the learning of the health index in the effective working period of the bearing, an abnormal threshold is obtained for the data of the bearing decay period. By fitting and analyzing the health index data during the recession period, a nonlinear state-space model consistent with the bearing recession process is constructed.

High Quality

Fast Deliver

Custom Service

Factory Price


After the model parameters are initialized with the Dempster-Shafer theory, the UKF filtering algorithm is used to update the degradation model parameters and the bearing remaining life. Prediction. Based on the test data of the bearing life cycle, the proposed method can better track the declining trend of bearing, and its performance is better than the remaining life prediction method based on the PF filter algorithm in the prediction of remaining life.

UKF Filter Algorithm

UKF is an algorithm for approximating the probability density function of the state. The algorithm is based on UT (unscented) transformation, uses the Kalman linear filter framework, and adopts the deterministic sampling method for sampling, which can effectively avoid the problem of particle degradation. . UT transformation is to extract a series of sampling points (sigma points) from the state prior distribution according to a certain sampling strategy so that the mean and covariance state distribution of these points is equal to the mean and covariance of the original state distribution; these sigma points are Substitute into the nonlinear function, and obtain the corresponding point set of nonlinear function value, and obtain the transformed mean and covariance through this point set. The most important thing in UT transformation is the sampling strategy of sigma points, the more commonly used is 2n+1 sigma point symmetric sampling. For an n-dimensional random variable with mean x- and covariance Px, an X matrix is generated, which consists of 2n+1 points, namely sigma points.
Calculation of bearing life

Custom solutions for your engineering needs.