Anomaly-Based Plant Control for Real-Time Stability Monitoring

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing control systems struggle to detect and respond to anomalies and cyber-attacks in real-time, especially when the plant or equipment physics diverge significantly from the assumed nominal range, leading to instability and failure, as they rely on fixed architectures and historical data, which are inadequate for unanticipated malfunctions.

Innovation Solution

The implementation of anomaly detection and anomaly-based control systems that use spectral correlations to compute stability radius measures and health indicators from runtime data, allowing for real-time identification of anomalies and adaptive control actions without requiring extensive historical data or system identification, enabling detection of unanticipated malfunctions and cyber-attacks.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If robust control theory is used to make the controller insensitive to friction disturbances, then the controller can handle a wider range of friction values, but the controller becomes confused and unstable when basic electro-mechanical physics changes or sensor malfunctions occur

Engineering Contradiction:
Improvefriction range handlingVSAvoidcontroller stability
Core Design Contradiction:
Adaptability or versatilityVSReliability

Solution Approach 1:

The controller dynamically adapts its architecture and parameters based on real-time anomaly detection. When anomalies are detected through spectral correlation analysis, the controller switches from a fixed robust control architecture to an adaptive architecture that adjusts gain schedules and control laws, allowing it to remain stable despite changes in electro-mechanical physics or sensor malfunctions.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system implements continuous feedback through anomaly detection that monitors spectral correlations between control inputs and plant responses. This feedback mechanism identifies when the plant behavior diverges from expected patterns, triggering adaptive control actions that restore stability and performance.

Inventive Principle:
Principle #23Feedback

2Device complexity

If fixed controller architectures are used, then the controller is simple to implement, but it cannot detect or respond to unanticipated malfunctions or cyber-attacks in real-time

Engineering Contradiction:
Improvecontroller architectureVSAvoidanomaly detection capability
Core Design Contradiction:
Device complexityVSReliability

Solution Approach 1:

The system pre-computes and stores spectral correlation patterns and stability radius thresholds during the design phase. These pre-computed references enable real-time anomaly detection without requiring complex online calculations, allowing the simple fixed controller architecture to gain enhanced reliability through pre-prepared diagnostic capabilities.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent introduces an intermediary anomaly detection layer that sits between the simple fixed controller and the plant. This intermediary continuously analyzes spectral correlations and stability radii, detecting anomalies that the simple controller cannot identify, thereby enhancing reliability without complicating the core control architecture.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Measurement precision

If extensive historical data collection is performed for system identification, then the controller can be more accurately tuned, but the computational complexity and data storage requirements increase significantly

Engineering Contradiction:
Improvesystem identification accuracyVSAvoiddata collection infrastructure
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The system performs partial system identification by focusing only on the essential spectral correlation patterns and stability radius thresholds needed for anomaly detection. Rather than collecting and analyzing extensive historical data for complete system identification, the patent uses minimal runtime data combined with pre-computed references, achieving sufficient accuracy while dramatically reducing computational and storage requirements.

Inventive Principle:
Principle #16Partial or excessive action

4Adaptability or versatility

If the controller uses stiff algorithms to be insensitive to friction disturbances, then the controller can operate across a wider friction range, but the controller may excite mechanical resonant modes when friction is very low or very high

Engineering Contradiction:
Improvefriction range operationVSAvoidmechanical resonance excitation
Core Design Contradiction:
Adaptability or versatilityVSObject-affected harmful factors

Solution Approach 1:

The controller dynamically changes its parameters including gain schedules and control laws based on detected friction conditions and anomaly indicators. When spectral analysis indicates abnormal friction conditions, the controller adjusts its parameters to avoid exciting resonant modes, allowing wide friction range operation without the harmful effects of resonance excitation.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP3759559B1Anomaly detection and anomaly-based control
Publication Date: 2022.11.30 WOODWARD INC
  • EP3759559B1 patent drawingFigure 1~2
  • EP3759559B1 patent drawingFigure 3
  • EP3759559B1 patent drawingFigure 3

AI summary

A plant control system includes a plant system and a control system controlling the plant system. Runtime conditions of an operating point of the plant control system are received. The runtime conditions include a runtime state of the plant system, a runtime output of the plant system, and a runtime control action applied to the plant system. Reference conditions of a reference point corresponding to the operating point are determined. Stability radius measures of a state difference, an output difference, and a control action difference are computed. One or more of an observability anomaly indicator, health observability indicator, tracking performance anomaly indicator, tracking performance health indicator, controllability anomaly indicator, and controllability health indicator are determined based on respective spectral correlations between two of the stability radius measure of the output difference, the stability radius measure of the state difference, and the stability radius measure of the control action difference.