Plant Diagnostics Using Self-Organizing Maps and Dynamic Time Warping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current monitoring methods for technical plants, especially in the process industry, are limited in their ability to detect deviations in manipulated and measured variables in an automated and reliable manner, particularly when fluctuations occur within fixed parameterized limits, and are complex to parameterize, especially in large plants.
Innovation Solution
A diagnostic device using a self-organizing map and the Dynamic Time Warping method to determine diagnostic statements about the operation of technical plants, allowing for automated monitoring of process variables and detecting anomalies without manual parameterization, by training on historical data to establish tolerance ranges and evaluate temporal sequences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If automated monitoring is implemented using fixed parameterized limits, then monitoring can be performed automatically, but the parameterization becomes very complex and deviations within parameterized limits cannot be detected
Solution Approach 1:
The system performs self-parameterization by automatically learning normal operation patterns from historical data without requiring manual parameter setup. The self-organizing map autonomously adapts to the specific plant's operational characteristics, eliminating the need for complex manual parameterization while maintaining automated monitoring.
Solution Approach 2:
The system pre-trains the self-organizing map using historical operational data before actual monitoring begins. This preliminary training phase allows the system to learn normal operation patterns in advance, so that during operation, only anomaly detection is required without complex real-time parameterization.
2Measurement precision
If user monitoring is performed with broad experience and high attentiveness, then monitoring accuracy can be maintained, but the user reaches their limits as the plant size increases
Solution Approach 1:
The patent replaces the mechanical human monitoring system with an automated evaluation device based on self-organizing maps. This substitution maintains high monitoring accuracy by objectively comparing current operation against learned normal patterns, while eliminating human limitations such as fatigue and capacity constraints.
Solution Approach 2:
The self-organizing map acts as an intermediary between raw operational data and human decision-making. It automatically processes and interprets complex multi-variable data, presenting only relevant anomalies to users, thus maintaining accuracy while reducing operational complexity.
3Reliability
If fixed limits are parameterized for monitoring, then automated detection is possible, but deviations occurring within parameterized limits cannot be detected
Solution Approach 1:
The system replaces static fixed limits with dynamic adaptive boundaries learned from historical data. The self-organizing map continuously adapts to normal operational variations, allowing the detection of deviations that fall within previously fixed limits but represent actual anomalies based on learned patterns.
Solution Approach 2:
The monitoring approach changes from using fixed parameterized limits to using dynamically learned operational patterns. The system transforms the parameter space by organizing multiple variables into a self-organizing map structure, enabling detection of deviations within traditional limits through pattern recognition rather than threshold comparison.
Data Source
AI summary
A diagnostic device for monitoring the operation of a technical plant with an automation system, wherein the diagnostic device includes a data memory in which at least one data set characterizing the operation of the plant with values of process variables can be stored, and an evaluation device, where the diagnostic device is characterized in that the evaluation device is configured to determine a diagnostic statement about the operation of the technical plant based on the data set and at least one self-organizing map and based on a program for controlling a sequence during the operation of the technical plant with repeatedly traversed step sequences via a Dynamic Time Warping method.