Digital Twin Fault Simulation for Automated Error Cause Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In automated systems, identifying the causes of errors is challenging due to temporal and local error propagation, making it difficult to trace and localize errors, especially with the combinatorial explosion of possible system states and behaviors, which existing methods like FMEA and RCA struggle to simulate comprehensively.
Innovation Solution
A method involving continuous detection of real system status, initialization of a digital twin with the most recent error-free state, simulation up to the error time with assumed faulty elements, and iterative comparison with real system states to identify faulty elements, using sensors and a control unit to automate this process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If comprehensive error simulation is performed using digital twins to identify all possible fault causes, then error identification completeness is improved, but computational complexity and analysis time increase exponentially due to the combinatorial explosion of possible system states
Solution Approach 1:
The patent applies preliminary action by pre-initializing the digital twin with the most recent error-free system state before simulation begins. This preparation step establishes a known good baseline, allowing the simulation to efficiently trace forward from a verified starting point rather than exploring all possible states from scratch, thus reducing computational complexity while maintaining comprehensive error identification
Solution Approach 2:
The patent uses copying by creating a virtual replica (digital twin) of the physical automated system. This copy allows comprehensive error simulation and analysis to be performed on the virtual model without affecting the real system, enabling thorough error identification while avoiding the computational burden of analyzing every possible physical system state directly
2Ease of manufacture
If traditional FMEA or RCA methods are used to model and analyze error causes, then analysis structure is improved, but the ability to identify previously unknown system states or errors deteriorates due to limited modeling contexts
Solution Approach 1:
The patent applies dynamics by using a simulation-based approach that can dynamically adapt to any system state rather than being constrained by pre-defined static models. The digital twin simulation can explore previously unknown system states and error conditions that were not anticipated in traditional FMEA models, thereby improving adaptability while maintaining structured analysis through the simulation framework
Solution Approach 2:
The patent applies universality by creating a digital twin that can represent and simulate any system state or error condition of the automated system, making the analysis method universally applicable to unknown or previously unmodeled errors. This multi-functional simulation capability allows the same approach to handle both known and unknown error types, enhancing versatility while preserving analytical structure
3Reliability
If simulation is performed from the beginning of operation to identify error causes, then analysis completeness is improved, but analysis time increases unnecessarily since the error cause may be in the past and no longer identifiable
Solution Approach 1:
The patent applies preliminary action by pre-identifying and recording the most recent error-free system state before the error occurred. This preparation allows the simulation to start from a known good state and trace forward only to the point where the error manifests, rather than simulating from the beginning of operation, thus reducing analysis time while maintaining completeness by capturing the critical transition point
Solution Approach 2:
The patent applies skipping by rapidly simulating forward from the identified error-free state to the error time point without dwelling on intermediate states. This allows the system to quickly traverse through the critical period where the error cause becomes apparent, reducing analysis time by focusing computational resources on the relevant time window rather than processing the entire operational history
Data Source
Figure 1
Figure 2
AI summary
The present invention relates to a method and to a device for identifying causes of faults in automated systems, and to an automated system comprising the device for identifying causes of faults in automated systems. In a digital twin of the automated system, at least one element of the digital twin is assumed to be faulty and then simulated using the digital twin until a fault time. At least one faulty element of the automated system is identified as the cause of a fault based on the at least one element assumed to be faulty.