Sensor Signal Anomaly Detection via Reference Model Learning

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

The maintenance and monitoring of critical equipment, such as power generating systems, are challenging due to numerous root causes of out-of-normal operational behavior, making exhaustive characterization difficult and requiring highly trained personnel, leading to high costs and time consumption.

Innovation Solution

A method and system that utilize sensors to capture signals and process them with a processor to define a reference model of normal operation in a first feature space, with the ability to analyze signals in orthogonal second feature spaces to detect out-of-normal conditions, employing techniques like principal component analysis or mutual interdependence analysis, allowing for automatic detection and alerting of equipment anomalies without requiring specialized staff.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional monitoring methods are used with highly trained personnel, then measurement precision and reliability are improved, but loss of time and loss of substance increase

Engineering Contradiction:

Improvemonitoring precisionVSAvoidtime consumption

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system performs self-diagnosis and self-monitoring by automatically comparing sensor data against learned normal signatures, eliminating the need for continuous human expert intervention while maintaining high measurement precision

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

Manual monitoring by trained personnel is replaced with an automated computer-based system that uses machine learning algorithms and signal processing to detect anomalies, thereby reducing time consumption while preserving monitoring accuracy

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If exhaustive characterization of failure mechanisms is attempted, then measurement precision is improved, but device complexity and loss of substance increase

Engineering Contradiction:

Improvecharacterization precisionVSAvoidsystem complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

Instead of characterizing all possible failure modes, the system inverts the approach by learning what normal operation looks like and detecting deviations from this baseline, thereby achieving comprehensive monitoring without exhaustive failure mechanism characterization

Inventive Principle:

Principle #13The other way round (Inversion)

Solution Approach 2:

The system creates a universal monitoring framework that can detect various types of anomalies (vibration, temperature, pressure, acoustic emissions) using a single integrated approach, reducing device complexity while maintaining comprehensive characterization capability

Inventive Principle:

Principle #6Universality (Multi-functionality)

3Reliability

If multiple sensors and analysis methods are deployed, then reliability is improved, but device complexity increases

Engineering Contradiction:

Improvemonitoring reliabilityVSAvoidsystem complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The system merges multiple sensor types (vibration, temperature, pressure, acoustic emissions) and analysis methods into a unified monitoring platform that uses a single learning algorithm to process all data streams, thereby improving reliability through multi-parameter monitoring while managing complexity through integration

Inventive Principle:

Principle #5Merging (Combining)

4Loss of time

If automated monitoring systems are implemented, then loss of time and loss of substance are reduced, but measurement precision may deteriorate

Engineering Contradiction:

Improveresponse timeVSAvoidanomaly detection accuracy

Core Design Contradiction:

Loss of timeVSMeasurement precision

Solution Approach 1:

The system performs preliminary learning during normal operation to establish baseline signatures, enabling rapid real-time detection of deviations without sacrificing measurement precision, as the automated system is pre-trained to recognize subtle anomalies

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS9443201B2Systems and methods for learning of normal sensor signatures, condition monitoring and diagnosis

Publication Date: 2016.09.13 SIEMENS AG

AI summary

Systems and methods to monitor a signal from an apparatus are disclosed. A feature extracted from the signal is automatically defined. Signals are received over a period of time wherein the apparatus is in a normal operational mode. Features are classified in a learning mode and are applied to create a reference model that defines a within-normal operational mode. In a testing mode a signal generated by the apparatus is received, a feature is extracted and classified. Instantaneous data generated in operational mode by the apparatus is classified by the system as abnormal if it does not lie within boundaries of the reference model or contains information/structure in an orthogonal subspace. A learned reference model is augmented by a user or automatically. In one illustrative example the apparatus is a power generation equipment and the signal is an acoustic signal.