Graph-Based Correlation Search for Industrial Process Control

Solution Overview

Problem

Existing technologies face challenges in efficiently finding significant multi-variate correlations within large datasets generated during complex manufacturing processes, such as semiconductor manufacturing, due to the computational intractability of analyzing large numbers of time series data streams.

Innovation Solution

A computer-implemented method using graph-based techniques to efficiently search for significant multi-variate correlations by generating pairwise correlation values, determining a variable of interest, and performing a graph search to filter variables using a heuristic based on pairwise correlation values, thereby supporting industrial process control.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If traditional statistical methods are used to find univariate correlations, then computational speed is fast and efficient, but the ability to detect nonfunctional relationships is limited

Engineering Contradiction:

Improvecomputational speedVSAvoidcorrelation detection accuracy

Core Design Contradiction:

SpeedVSMeasurement precision

Solution Approach 1:

The patent segments the correlation search process into two distinct phases: (1) a fast univariate correlation screening phase using traditional statistical methods to quickly eliminate obviously uncorrelated variable pairs, and (2) a more thorough multivariate correlation analysis phase using modern statistical techniques on the reduced subset of candidate variables. This segmentation allows the system to achieve both computational efficiency and accurate detection of nonfunctional relationships.

Inventive Principle:

Principle #1Segmentation

2Measurement precision

If modern statistical techniques are used to detect any relationship including nonfunctional ones, then correlation detection accuracy improves, but computational resources required increase significantly

Engineering Contradiction:

Improvecorrelation detection accuracyVSAvoidcomputational resource consumption

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent segments the correlation search process into two distinct phases: (1) a fast univariate correlation screening phase using traditional statistical methods to quickly eliminate obviously uncorrelated variable pairs, and (2) a more thorough multivariate correlation analysis phase using modern statistical techniques on the reduced subset of candidate variables. This segmentation allows the system to achieve both computational efficiency and accurate detection of nonfunctional relationships.

Inventive Principle:

Principle #1Segmentation

3Loss of information

If all pairs of variables are analyzed for correlation, then comprehensive correlation coverage is achieved, but computational feasibility becomes intractable with large numbers of variables

Engineering Contradiction:

Improvecorrelation coverage completenessVSAvoidcomputational efficiency

Core Design Contradiction:

Loss of informationVSProductivity

Solution Approach 1:

The patent segments the correlation search process into two distinct phases: (1) a fast univariate correlation screening phase using traditional statistical methods to quickly eliminate obviously uncorrelated variable pairs, and (2) a more thorough multivariate correlation analysis phase using modern statistical techniques on the reduced subset of candidate variables. This segmentation allows the system to achieve both computational efficiency and accurate detection of nonfunctional relationships.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent applies partial action by not analyzing all possible variable pairs with computationally intensive modern statistical techniques. Instead, it performs exhaustive univariate correlation analysis on all pairs first, then applies the more resource-intensive multivariate analysis only to the subset of pairs that show significant univariate correlation. This partial application of the more demanding method maintains comprehensive coverage of likely correlations while avoiding the computational intractability of applying it to all pairs.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS20250036089A1Accelerating search for multi-variate correlations for industrial process control

Publication Date: 2025.01.30 DELTA DESIGN INC

AI summary

In some embodiments, a computer-implemented method of improving industrial process control by finding significant multi-variate correlations within a plurality of variables representing sensor data is provided. A computing system obtains time series data streams for the plurality of variables. The computing system generates pairwise correlation values between the variables of the plurality of variables. The computing system determines a variable of interest from the plurality of variables, and performs a graph search to determine one or more significant multi-variate correlations between variables from the plurality of variables and the variable of interest. Variables are filtered from the graph search using a heuristic based on the pairwise correlation values. The multi-variate correlations are provided to support the industrial process control.