Fault Characterization via Healthy Model Simulation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for characterizing failures in complex industrial systems are inadequate, as they rely on limited sensor data and require extensive expert analysis, leading to delayed maintenance and potential false alarms, and are not applicable to new systems without a measurement history database.
Innovation Solution
A method that models the system using a healthy model to simulate internal physical quantities and external boundary conditions, creating a failure matrix to identify and characterize failures by comparing measured and simulated data, using pseudo-random iterations to refine failure vector probabilities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If learning algorithms are used to characterize failures, then failure detection capability is improved, but the method becomes difficult to apply to complex systems and generates false alarms
Solution Approach 1:
The patent introduces a virtual model as an intermediary between the physical system and the analysis tools. This virtual model simulates system behavior under various failure conditions, enabling complex failure characterization without requiring direct complex learning algorithms on the actual system data, thus resolving the contradiction between detection capability and system applicability
Solution Approach 2:
The patent creates a virtual copy (digital twin) of the physical system that replicates its behavior. This copy is used to generate synthetic failure data and characterize failures without needing to apply complex learning algorithms directly to the complex physical system, thereby improving detection capability while avoiding the limitations of applying such algorithms to complex real systems
2Measurement precision
If expert analysis is used to characterize failures, then characterization precision is improved, but maintenance time is delayed
Solution Approach 1:
The system performs self-diagnosis by automatically comparing sensor measurements with virtual model predictions and characterizing failures through algorithmic analysis. This eliminates the need for time-consuming expert analysis while maintaining high characterization precision, thus resolving the contradiction between precision and time loss
Solution Approach 2:
The patent replaces the mechanical process of expert analysis with an automated computational system. The virtual model and comparison algorithms automatically characterize failures by processing sensor data, substituting human expert intervention with an automated system that achieves both high precision and rapid response
3Reliability
If learning algorithms are used, then failure characterization is improved, but the system requires prior measurement history database
Solution Approach 1:
The patent performs preliminary actions by pre-building a virtual model of the system with its expected behavior characteristics before actual operation. This virtual model enables immediate failure characterization capability from the start of system operation, eliminating the need for a prior measurement history database and allowing new systems to be applicable immediately
Data Source
Figure 1
Figure 2
AI summary
The invention relates to a method for characterizing one or more faults in a system grouping together a plurality of internal physical quantities and delimited by a plurality of limit physical quantities, the system being modelled by a healthy model establishing relationships linking said internal physical quantities with one another and with the limit physical quantities in the absence of a fault, a fault being defined as a negative change in the relationships linking said internal physical quantities with one another and with the limit physical quantities with respect to the healthy model, wherein a fault is characterized by recording a number of iterations having involved said fault in a series of iterations involving a fault matrix.