Machine Operating Models for Adaptive Anomaly Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Complex systems like power plants and factories face inefficiencies due to unpredictable machine behavior and sensor degradation, requiring accurate maintenance decisions that are often challenging for human engineers to make, especially in large and varied environments.

Innovation Solution

A computer-implemented method that generates and updates machine operating models based on past data to predict behavior across different states, using generic definitions and diagnostic logic to continuously adapt and detect anomalies, enabling dynamic management of machine health.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If human maintenance engineers make maintenance decisions based on sensor data and expertise, then maintenance decisions can be made with human judgment and adaptability, but the accuracy and consistency of decisions deteriorate due to human limitations and variability

Engineering Contradiction:
Improvehuman judgment adaptabilityVSAvoiddecision accuracy
Core Design Contradiction:
Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The system enables machines to monitor and diagnose their own operational states through embedded sensors and processing units that continuously collect and analyze operational data, reducing dependence on human maintenance engineers for real-time decision-making while maintaining adaptability through machine learning algorithms

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent replaces human mechanical judgment processes with automated electronic systems including sensors, processing units, and machine learning models that objectively analyze operational data and predict maintenance needs, eliminating human variability while preserving adaptability through continuous learning

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Reliability

If sensors continuously monitor machine operational behavior to detect anomalies, then real-time detection capability is improved, but measurement precision deteriorates due to sensor degradation and failure

Engineering Contradiction:
Improveanomaly detection capabilityVSAvoidsensor measurement accuracy
Core Design Contradiction:
ReliabilityVSMeasurement precision

Solution Approach 1:

The system implements feedback mechanisms where machine learning models continuously learn from historical sensor data and operational outcomes, adjusting their predictions and thresholds to compensate for sensor degradation over time, thereby maintaining reliable anomaly detection despite deteriorating sensor precision

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The patent dynamically adjusts monitoring parameters and thresholds based on learned patterns from historical data, adapting the sensitivity and focus of monitoring to account for sensor degradation, changing which parameters are monitored most closely as the system evolves

Inventive Principle:
Principle #35Parameter changes

3Ease of manufacture

If maintenance schedules are based on generic manuals and specifications to ensure consistent maintenance practices, then standardization is improved, but productivity deteriorates due to over-estimation or under-estimation of maintenance frequency

Engineering Contradiction:
Improvemaintenance standardizationVSAvoidsystem efficiency
Core Design Contradiction:
Ease of manufactureVSProductivity

Solution Approach 1:

The system transitions from static, fixed maintenance schedules to dynamic, adaptive maintenance planning where machine learning models continuously predict actual maintenance needs based on real-time operational data, allowing maintenance frequency to automatically adjust to actual machine conditions rather than following rigid generic schedules

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

Machines autonomously generate their own maintenance schedules based on their actual operational patterns and predicted degradation trajectories, eliminating dependence on generic manufacturer recommendations and enabling each machine to optimize its own maintenance timing for maximum efficiency

Inventive Principle:
Principle #25Self-service

4Measurement precision

If complex monitoring systems are implemented to track multiple machine components to improve detection accuracy, then anomaly detection capability is improved, but device complexity increases making the system harder to manage

Engineering Contradiction:
Improvebehavior detection accuracyVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent combines multiple sensor inputs and monitoring functions into unified machine learning models that process diverse operational data together, detecting anomalies through integrated analysis rather than separate monitoring of each component, thereby maintaining high detection accuracy while reducing the complexity of individual monitoring subsystems

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The system employs universal machine learning algorithms and processing frameworks that can handle multiple types of sensors and machine components through a single cohesive architecture, enabling the system to detect anomalies across diverse components without requiring separate complex monitoring systems for each

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentEP3074829B1Learning expected operational behavior of machines from generic definitions and past behavior
Publication Date: 2023.05.31 FALKONRY INC
  • EP3074829B1 patent drawingFigure 1
  • EP3074829B1 patent drawingFigure 2
  • EP3074829B1 patent drawingFigure 3

AI summary

In an embodiment, a data processing method comprises storing one or more generic machine operating definitions, wherein each of the generic machine operating definitions describes expected operational behavior of one or more types of machines during one or more operating states; analyzing operating data that describes past operation of a plurality of machines of a plurality of types; based at least in part on the operating data and the one or more generic machine operating definitions, generating and storing one or more machine operating models that describe expected operational behavior corresponding to a plurality of operating states of the plurality of machines; wherein the one or more machine operating models comprise a plurality of data patterns, wherein each of the data patterns is associated with a different set of one or more operating states of one or more machines; wherein the method is performed by one or more computing devices.