Failure Mode Relative Likelihood Algorithm for Fault Diagnosis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Ambiguity in determining the root cause of failures in monitored systems due to overlapping evidence, leading to uncertainty in fault reasoning, often requiring complex and expensive software to distinguish between contributing system components.
Innovation Solution
The failure-mode-relative-likelihood algorithm computes the relative probability of failure modes based on evidence observations, using a reference model of the system that incorporates a Noisy-OR model and naïve Bayesian reasoning, accounting for false alarms and unknown factors, to determine the likelihood of failure modes in a monitored system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If complex software is used to determine root cause failures, then measurement precision is improved, but device complexity increases and cost increases
Solution Approach 1:
The patent replaces complex, expensive software with a simple probabilistic algorithm that uses basic mathematical operations (multiplication, addition, division) to compute failure likelihoods. The solution uses inexpensive monitor data and straightforward probability calculations instead of sophisticated diagnostic software.
Solution Approach 2:
The patent transforms the fault diagnosis problem from a complex logical analysis into a probabilistic parameter calculation. By changing the approach from qualitative fault tree analysis to quantitative probability computation using likelihood ratios, the system achieves accurate root cause identification with simpler methods.
2Measurement precision
If complex software is used to determine root cause failures, then measurement precision is improved, but loss of energy increases
Solution Approach 1:
The patent uses computationally inexpensive probability calculations that require minimal processing power. The algorithm computes likelihood ratios using basic arithmetic operations on monitor data, avoiding the energy-intensive computations required by complex diagnostic software.
3Reliability
If monitors are connected in many-to-many relationship to failure modes, then reliability is improved, but device complexity increases
Solution Approach 1:
The patent handles the many-to-many monitor-failure mode relationships by transforming the complexity into probabilistic parameter calculations. Instead of managing complex logical relationships, the system uses likelihood ratios and probability computations to naturally handle multiple monitors indicating multiple failure modes, simplifying the analysis while maintaining reliability.
Data Source
AI summary
A method for determining relative likelihood of a failure mode is provided. The method comprises receiving evidence observations of a monitored system from monitors connected in a many-to-many relationship to the failure modes, generating a fault condition including states of all failure modes that are connected to the monitors, and computing a relative probability of failure for each failure mode. The fault condition is generated for a reference model of the monitored system and is based on the received evidence observations. The relative probability of failure for each failure mode is based on a false alarm probability, a detection probability, and a ratio of prior probabilities of a candidate hypothesis to a null hypothesis of no active failure mode.