Lubricant Condition Assessment Across Multiple Operating Modes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing liquid lubrication systems face challenges in accurately assessing the condition of lubricant fluids due to changes in operating intensity, leading to concealed drifts and potential system damage, as current methods struggle to differentiate between changes caused by operation mode shifts and other factors, resulting in imprecise predictions and delayed detection of issues.
Innovation Solution
A method for liquid condition assessment in a lubrication system that involves measuring monitored data, determining the current operating mode, selecting a corresponding mathematical model, and comparing it with simulated data to identify changes and drifts, using multiple formulas to account for different operation modes and variables, thereby providing a normal distribution for precise analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If statistical analysis with logarithmic transformation is used to improve data distribution, then the data resembles a normal distribution more, but drifts from one island to another are concealed and not realised until passed
Solution Approach 1:
The patent segments the continuous operational data into distinct operating mode clusters (islands) using unsupervised machine learning algorithms. Each cluster represents a specific operating mode with its own characteristics. By segmenting the data this way, the system can track transitions between modes and detect drifts within and between modes, resolving the contradiction between achieving normal distribution and detecting drifts.
2Measurement precision
If multiple operating modes are monitored with separate models, then each mode can be analysed accurately, but the system complexity increases and drift detection becomes more difficult
Solution Approach 1:
The patent employs a universal machine learning framework that automatically adapts to different operating modes without requiring separate manual models for each mode. The unsupervised learning algorithm dynamically identifies and tracks multiple clusters (operating modes) within a single unified system, enabling accurate mode-specific analysis while maintaining manageable system complexity through automation.
3Ease of operation
If traditional threshold-based monitoring is used, then the system is simple to operate, but changes in operation intensity are misinterpreted as system drifts leading to false warnings
Solution Approach 1:
The patent transitions from static threshold-based monitoring to dynamic, adaptive monitoring using machine learning models that automatically adjust to changing operating conditions. The system dynamically identifies normal variations associated with different operating modes and distinguishes these from actual system drifts, significantly reducing false warnings while maintaining operational simplicity through automated adaptation.
Data Source
AI summary
The present disclosure relates to a method and an apparatus for liquid condition assessment in a liquid lubrication system for a work system operable in at least two different modes, the method comprising measuring monitored liquid condition data, determining a current operating mode of the work system, then based on the current operating mode, selecting a corresponding model, modeling corresponding simulated liquid condition data based on the model, assessing the liquid condition by comparing the monitored liquid condition data with the simulated liquid condition data, and providing a normally distributed data set for the liquid condition assessment.