Equation-Level Diagnostic Error Messages for Circuit Simulation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional mechanisms for diagnosing illegal configurations in circuit simulation, especially in multi-domain systems, face challenges due to the inability to properly categorize components into predefined classes, leading to inaccurate diagnoses and failure to detect all types of illegal configurations.
Innovation Solution
The approach involves converting system model components into numerical equations, performing numerical analysis to identify illegal configurations, and generating equation-level diagnostic error messages without requiring pre-categorization of components, using techniques like sparse QR decomposition and reduced-row-echelon form to analyze the equations and detect dependencies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional structural analysis is used to diagnose illegal configurations, then the analysis process is simple and straightforward, but it fails to detect all types of illegal configurations especially in multi-domain systems
Solution Approach 1:
The patent replaces conventional structural analysis methods with numerical analysis of equations. Instead of using traditional structural approaches that categorize components into five classes, the invention uses numerical evaluation of the Jacobian matrix to detect illegal configurations. This substitution enables detection of all types of illegal configurations including those in multi-domain systems while maintaining computational efficiency through systematic numerical methods.
2Reliability
If pre-categorization of components into five classes is performed, then the analysis structure is organized and systematic, but it becomes impossible or awkward for multi-domain systems and significantly reduces diagnosis accuracy
Solution Approach 1:
The patent implements a universal analysis approach that works across all domain types without requiring pre-categorization. The numerical analysis method evaluates equations and their Jacobian matrices regardless of whether components belong to electrical, mechanical, thermal, or other domains. This universal approach eliminates the need for awkward categorization choices in multi-domain systems while maintaining high diagnosis accuracy through equation-based analysis.
3Loss of information
If conventional diagnostic mechanisms are used, then the implementation is straightforward and requires minimal computational resources, but the error messages fail to identify all components involved in illegal configurations
Solution Approach 1:
The patent employs feedback mechanisms through iterative numerical analysis of the Jacobian matrix. The system automatically evaluates equations, computes the Jacobian, detects singularities or near-singularities, and traces back to identify all components involved in illegal configurations. This automated feedback loop ensures complete error information is captured and presented to users without manual intervention, significantly improving error message completeness.
Data Source
AI summary
A mechanism for providing equation-level diagnostic error messages for system models undergoing circuit simulations is discussed. The components in a model of a system being simulated are converted into multiple numerical equations where each equation corresponds to a component in the system being simulated or a topology equation for the system model. Each numerical equation is numerically analyzed in order to identify illegal configurations in the system. Upon detection of an error, an error message listing the components associated with the illegal configuration is generated for the user.