Signal Analysis Shell Manifold for Adaptive Alarm Bounds

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current methods for quantitative analysis of signal-related measurements are limited in handling moving signature conditions and require manual adjustment of alarm levels, leading to issues with false positives and false negatives, and lack continuous self-adjustment of sensitivity.

Innovation Solution

A computerized method that projects signal-related measurements into a multidimensional space, computes multidimensional statistics, and quantifies matching likelihoods based on distances to a shell manifold derived from these statistics, allowing for continuous adjustment of alarm bounds and sensitivity.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If alarm levels are set high to avoid false positive indications, then false positive rate is reduced, but sensitivity decreases and false negative rate increases

Engineering Contradiction:

Improvefalse positive rateVSAvoidsensitivity

Core Design Contradiction:

ReliabilityVSMeasurement precision

Solution Approach 1:

The patent implements dynamic alarm bounds that automatically adjust to changing process conditions and signature variations. Instead of fixed high alarm levels, the system continuously adapts the alarm thresholds based on learned process behavior, enabling both high reliability by reducing false positives and maintained sensitivity by adjusting bounds to actual process variability rather than using conservative fixed levels

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system performs self-adjustment of alarm levels through automatic signature recognition and adaptation. The monitoring algorithm autonomously learns process signatures and adjusts alarm bounds without manual intervention, eliminating the need for operators to manually set high alarm levels while maintaining both reliability and sensitivity through continuous self-optimization

Inventive Principle:

Principle #25Self-service

2Adaptability or versatility

If manual adjustment of alarm levels is used, then alarm bounds can be customized, but adjustment accuracy is poor and robustness decreases

Engineering Contradiction:

Improvealarm bound customizationVSAvoidalarm level accuracy

Core Design Contradiction:

Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The system automatically performs signature recognition and alarm bound adjustment without requiring manual operator input. The algorithm autonomously analyzes process data, identifies signatures, and sets optimal alarm levels, achieving both high adaptability to different processes and high precision in alarm level setting through automated statistical analysis and machine learning techniques

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The system implements continuous monitoring and automatic adaptation where alarm bounds are dynamically adjusted based on feedback from ongoing process data analysis. The system learns from process behavior patterns and continuously refines alarm levels, achieving accurate and robust alarm settings that adapt to changing conditions without manual re-adjustment

Inventive Principle:

Principle #23Feedback

3Device complexity

If only the greatest excess is considered for alarms, then alarm simplicity is maintained, but sensitivity to multiple small deviations is lost

Engineering Contradiction:

Improvealarm evaluation simplicityVSAvoiddeviation detection sensitivity

Core Design Contradiction:

Device complexityVSMeasurement precision

Solution Approach 1:

The patent transitions from one-dimensional scalar alarm evaluation to multidimensional vector-based signature analysis. Instead of considering only the magnitude of individual deviations, the system analyzes patterns across multiple dimensions simultaneously, enabling detection of multiple small deviations that collectively indicate process anomalies while maintaining computational efficiency through vector operations

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

4Ease of manufacture

If fixed alarm levels are used, then implementation is simple, but robustness and sensitivity are reduced

Engineering Contradiction:

Improveimplementation simplicityVSAvoidmonitoring robustness

Core Design Contradiction:

Ease of manufactureVSReliability

Solution Approach 1:

The system implements dynamic alarm bounds that automatically adapt to changing process conditions through continuous signature recognition and statistical analysis. This maintains implementation simplicity by using automated algorithms while dramatically improving robustness and sensitivity through adaptive adjustment to actual process variability rather than relying on conservative fixed thresholds

Inventive Principle:

Principle #15Dynamics

Data Source

PatentUS10902088B2Quantitative analysis of signal related measurements for trending and pattern recognition

Publication Date: 2021.01.26 HYDRO QUEBEC CORP

AI summary

A computerized method for quantitative analysis of signal related measurements, performed with one or more processors, is disclosed. An estimated signature typifying a characteristic feature of the signal related measurements is produced. Multidimensional statistics on the signal related measurements are computed in a multidimensional space with respect to the estimated signature. Matching likelihoods of the signal related measurements are quantified based on distances of the signal related measurements with respect to a shell manifold derived from the multidimensional statistics and enveloping a signature manifold in the multidimensional space. Multidimensional statistics on the estimated signature and trending and pattern recognition are also possible from the signal multidimensional projection.