Detecting Influential Components in Electrical Circuit Netlists
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting parameter variations in electrical circuits are computationally expensive and inefficient, especially for larger circuits, as they require significant computational power to identify components with the strongest influence on circuit performance.
Innovation Solution
A computer-implemented method that uses Monte Carlo simulation and iterative orthogonal matching pursuit, incorporating topological patterns to analyze the electrical circuit's topology and model parameter variations, thereby reducing computational power by accounting for dependencies between interconnected components.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If Monte Carlo simulation is used to detect parameter variations in electrical circuits, then measurement precision is improved, but use of energy increases due to significant computational power requirements
Solution Approach 1:
The patent segments the circuit analysis by identifying and focusing only on influential components rather than analyzing all components uniformly. The method divides the computational task into identifying key components first, then analyzing only those components in detail, thereby reducing overall computational energy while maintaining detection precision.
Solution Approach 2:
The patent changes the approach from exhaustive simulation of all parameters to selective analysis based on influence metrics. By transforming the problem from analyzing all component parameters to analyzing only influential parameters, the method reduces computational energy consumption while preserving measurement precision through targeted Monte Carlo simulation on selected components.
2Measurement precision
If exhaustive simulation of all component parameters is performed, then measurement precision is improved, but productivity deteriorates due to computational expense
Solution Approach 1:
The methodology segments the circuit into influential and non-influential components, performing exhaustive simulation only on the influential subset. This segmentation maintains measurement precision for critical components while dramatically improving overall productivity by avoiding unnecessary simulations of non-critical components.
Solution Approach 2:
The patent applies partial action by performing complete Monte Carlo simulation only on influential components rather than all components. This partial exhaustive approach achieves sufficient measurement precision for circuit performance prediction while significantly improving productivity through reduced computational scope.
3Measurement precision
If Monte Carlo sampling is applied to larger electrical circuits, then measurement precision is maintained, but device complexity increases due to prohibitive computational requirements
Solution Approach 1:
The patent reduces device complexity by segmenting the large circuit into influential components that require detailed analysis and non-influential components that can be analyzed coarsely or excluded. This segmentation maintains measurement precision for robustness estimation while reducing the apparent complexity of the computational system.
Solution Approach 2:
The methodology changes the analysis parameters from comprehensive simulation of all circuit elements to selective simulation based on influence metrics. This parameter change maintains measurement precision by focusing computational resources on critical parameters while reducing overall device complexity through decreased computational burden.
Data Source
AI summary
A computer-implemented method and an electronic device for detecting, in an electrical circuit with electrical components (Mg) subject to variations (δ) of their model parameters (xj), those components (Mg*) the model parameter variations (δ) of which have the strongest influence on a performance (yn) of the circuit, comprising: providing a topology (Q) of the circuit and the model parameters (xj) of all components (Mg) therein; determining, therefrom, topological patterns (Pk) of interconnected components (Mg); generating variation samples (vn), each comprising a different set of candidate variations (x′j) of the model parameters (xj); calculating, for each variation sample (vn), the circuit's performance (yn) and a deviation from a standard performance and forming a deviation vector (yD) therefrom; and using the variation samples (vn), the deviation vector (yD) and the topological patterns (Pk) in a regression model for detecting the most influential components (Mg*).