Machine Signature Derivation Using Multi-Sensor Normal Condition Detection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing industrial monitoring systems struggle to accurately identify normal operating conditions in industrial operations due to the complexity of sensor data from multiple sources, often failing to differentiate between normal and abnormal conditions, especially when rhythmic patterns are present.

Innovation Solution

A method and system utilizing multiple sensor types, including primary and secondary sensors, in conjunction with a machine learning algorithm to determine normal operating conditions and derive a primary sensor signature, which helps identify abnormal conditions by correlating data from various sensors and filtering out transient information.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If multiple sensor types are used to monitor industrial operations, then measurement precision and reliability improve, but device complexity and difficulty of detecting normal vs abnormal conditions worsen

Engineering Contradiction:

Improveaccuracy of identifying normal operating conditionsVSAvoidcomplexity of sensor data from multiple sources

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the sensor data processing by separating primary sensors (which detect the parameter of interest) from secondary sensors (which detect transient conditions). The system processes secondary sensor data first to identify transient periods, then uses this information to selectively process primary sensor data, dividing the complex multi-sensor data into manageable segments that can be analyzed independently and then integrated.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces an intermediary processing layer that correlates data from multiple sensor types before final analysis. This intermediary layer integrates information from primary and secondary sensors, using the secondary sensor data as a mediator to filter and contextualize the primary sensor data, making the overall system more manageable despite the complexity of multiple sensor sources.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Reliability

If multiple sensor types are used to monitor industrial operations, then reliability of condition identification improves, but difficulty of detecting and measuring normal vs abnormal conditions worsens

Engineering Contradiction:

Improveaccuracy of condition identificationVSAvoiddifficulty of differentiating normal and abnormal conditions

Core Design Contradiction:

ReliabilityVSDifficulty of detecting and measuring

Solution Approach 1:

The patent performs preliminary action by using secondary sensors to identify transient periods before analyzing primary sensor data. The system proactively detects transient conditions using secondary sensors, then uses this advance knowledge to filter or exclude corresponding primary sensor data during signature development, preventing transient artifacts from contaminating the normal operating signature before the main analysis occurs.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent implements feedback by using secondary sensor information to continuously adjust and refine the interpretation of primary sensor data. The secondary sensors provide feedback about transient conditions that helps the system distinguish between normal variations and abnormal conditions in the primary sensor data, improving the overall reliability of condition identification through iterative refinement.

Inventive Principle:

Principle #23Feedback

3Reliability

If secondary sensor information is used to validate primary sensor data, then false alerts are reduced, but loss of time in processing additional sensor data increases

Engineering Contradiction:

Improvereduction of false alertsVSAvoidtime for processing sensor data

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent extracts and removes transient data from the primary sensor signal during periods when secondary sensors detect transient conditions. By taking out these problematic data segments before signature development, the system prevents transient artifacts from creating false alerts, while minimizing processing time by excluding rather than extensively analyzing the removed portions.

Inventive Principle:

Principle #2Taking out (Extraction)

Data Source

PatentUS11636752B2Monitoring machine operation with different sensor types to identify typical operation for derivation of a signature

Publication Date: 2023.04.25 ROCKWELL AUTOMATION TECH INC

AI summary

A method for derivation of a machine signature includes receiving sensor information from a primary sensor, where the primary sensor is positioned to receive information from a portion of an industrial operation, and receiving sensor information from one or more secondary sensors. The secondary sensors are arranged to provide additional information about the industrial operation indicative of current operating conditions of the industrial operation. The method includes using the sensor information from the secondary sensors and machine learning to determine if the portion of the industrial operation is operating in a normal condition and, in response to determining that the portion of the industrial operation is operating normally, using sensor information from the primary sensor during the normal operating condition to derive a primary sensor signature for the sensor information from the primary sensor.