Interactive Diagnostic Modeling Evaluator for Automated Fault Matrix Repair
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional diagnostic systems for fault detection, isolation, and recovery in complex systems, such as those used in space missions, face challenges in ensuring high performance and accuracy due to manual processes that become infeasible for large-scale systems with numerous sensors and failure modes, leading to high false positive and negative detection rates, and lack of automated mechanisms for verifying diagnostic model accuracy.
Innovation Solution
The development of an interactive diagnostic modeling evaluator (i-DME) that iteratively repairs diagnostic units by computing diagnostic performance, proposing repairs, and re-evaluating based on performance metrics, allowing for automated verification and correction of diagnostic models, thereby improving diagnostic performance and reducing human error.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If manual diagnostic modeling is used for large-scale systems with numerous sensors and failure modes, then human modelers can understand and create diagnostic models, but the modeling process becomes infeasible and beyond human comprehension due to the complexity of n! potential fault interactions
Solution Approach 1:
The patent replaces manual human modeling processes with an automated computer-based system that uses algorithms to generate diagnostic models. The system automatically processes sensor data and failure mode information to create diagnostic matrices, eliminating the need for human modelers to manually analyze complex fault interactions.
Solution Approach 2:
The diagnostic system performs self-verification through automated mechanisms that check model accuracy and diagnostic performance without human intervention. The system automatically evaluates diagnostic utility measures and identifies areas for model improvement, enabling self-correction and continuous refinement of the diagnostic model.
2Reliability
If the diagnostic system model is developed manually to capture all failure modes, then comprehensive fault coverage is achieved, but the false positive and false negative detection rates become too high due to human error and inability to process large datasets
Solution Approach 1:
The patent implements automated feedback mechanisms where the diagnostic system continuously evaluates its own performance using diagnostic utility measures. The system compares actual diagnostic outcomes against expected results and automatically adjusts the diagnostic model to reduce false positives and false negatives, improving detection precision over time.
Solution Approach 2:
The system creates and evaluates multiple diagnostic model variations automatically, testing different configurations and selecting the optimal model. This automated model generation and testing process replaces manual model creation, ensuring comprehensive failure mode coverage while maintaining high detection precision through systematic evaluation.
3Measurement precision
If human modelers manually visualize and correct diagnostic models to improve performance, then some diagnostic accuracy can be improved, but the process becomes infeasible for large-scale systems with thousands of FMEA failure modes
Solution Approach 1:
The patent replaces manual model visualization and correction processes with automated computer-based algorithms that systematically analyze diagnostic model performance. The system automatically identifies model deficiencies and generates corrections, eliminating the time-consuming manual process of visualizing and correcting models for large-scale systems with thousands of failure modes.
Solution Approach 2:
The diagnostic system performs preliminary automated evaluation of model accuracy before deployment, identifying potential issues with false positives and false negatives in advance. This preliminary automated analysis prevents the need for time-consuming manual corrections after deployment, saving significant time in the overall model development process.
4Reliability
If the diagnostic model includes all potential failure mode interactions, then complete fault coverage is achieved, but the computational complexity and difficulty of verifying model accuracy becomes unmanageable
Solution Approach 1:
The patent implements automated self-verification mechanisms where the diagnostic system automatically checks its own model accuracy and completeness. The system uses algorithms to verify that all failure modes are properly captured and that diagnostic logic is correct, eliminating the need for manual verification of complex model interactions.
Solution Approach 2:
The system automatically generates test cases and verification data to validate the diagnostic model comprehensively. By creating simulated failure scenarios and comparing actual diagnostic outcomes against expected results, the system verifies model accuracy without requiring manual analysis of complex fault interactions.
Data Source
AI summary
Interactive diagnostic modeling evaluator (i-DME) for repairing a diagnostic matrix of a system. The diagnostic modeling evaluator cause steps to be performed comprising: evaluating a first diagnostic performance of a diagnostic unit using supervised data that includes one or more instances of a fault of a system; proposing a repair to the diagnostic unit based on the evaluated diagnostic performance; evaluating a second diagnostic performance of the diagnostic unit having the proposed repair; and upon determining that the second diagnostic performance meets a performance metric criterion, repairing the diagnostic unit according to the proposed repair.