Bearing Remaining Life Estimation Using Defect Severity Modeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Bearing failures due to factors like poor lubrication and contamination lead to unexpected downtime and safety risks, as existing monitoring systems are inadequate in estimating remaining useful life effectively.
Innovation Solution
A method and apparatus using real-time operational data and machine learning models to monitor defects in bearings, compute impact severity, and determine remaining useful life through a virtual model based on dynamic parameters like contact force and stress, facilitating continuous and accurate estimation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional bearing monitoring systems are used, then basic defect detection is achieved, but remaining useful life estimation is inadequate
Solution Approach 1:
The system performs preliminary actions by continuously collecting and storing operational data (vibration, temperature, load, speed) before actual failure occurs. This enables the machine learning models to be trained on historical data and make accurate predictions about remaining useful life, resolving the contradiction between measurement precision and reliability.
Solution Approach 2:
Machine learning models act as intermediaries between raw sensor data and remaining useful life estimation. The models process operational data, identify patterns indicating defect progression, and predict remaining life with high accuracy, thereby improving both measurement precision and prediction reliability simultaneously.
2Reliability
If continuous monitoring is implemented, then bearing health tracking is improved, but system complexity increases
Solution Approach 1:
The system achieves universality by using a multi-functional approach where machine learning models perform multiple tasks: detecting defects, tracking progression, estimating remaining useful life, and predicting failures. This consolidates multiple monitoring functions into a unified system, improving reliability without proportionally increasing complexity.
Solution Approach 2:
The system implements self-service through automated machine learning models that continuously analyze operational data and generate predictions without manual intervention. The models automatically adapt to different bearing conditions and update remaining useful life estimates, reducing the need for complex manual monitoring procedures while maintaining high reliability.
3Measurement precision
If machine learning models are used for time period determination, then detection accuracy is improved, but computational requirements increase
Solution Approach 1:
The system applies partial action by using machine learning models selectively for critical predictions (remaining useful life estimation and failure prediction) rather than continuous full-scale analysis. This approach maintains high detection accuracy for key parameters while reducing overall computational energy consumption compared to exhaustive analysis of all operational data.
Data Source
Figure 1A
Figure 1B
Figure 2A
AI summary
The present invention provides a system (100), apparatus (110) and method for estimating remaining useful life of a bearing. In one embodiment, the method comprises receiving a request for analysing a defect in the bearing from a source (115, 125, 130). The request comprises operational data associated with the bearing. The method comprises monitoring an impact of the defect on the bearing over a period of time. The method comprises determining a time period during which the impact of the defect on the bearing is higher than a threshold range, using a machine learning model. The method comprises computing a severity of the impact associated with the defect during the time period. The method comprises determining a remaining useful life of the bearing based on the severity and the operational data during the time period.