Rotating Equipment Diagnostics Using Sensor and Text Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current diagnostic tools for gas turbines lack the ability to automatically integrate and analyze both sensor data and textual information, leading to inefficient fault detection and diagnosis, requiring extensive manual parameterization and multi-disciplinary expertise.
Innovation Solution
A method combining Natural Language Processing (NLP) and Deep Learning to integrate textual information with sensor data, creating a unified representation for optimized diagnostics, reducing troubleshooting time and increasing technical responsiveness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If manual diagnostic methods with free text documentation are used, then ease of operation for technicians is improved, but information sharing and knowledge reuse deteriorate
Solution Approach 1:
The system creates structured copies of diagnostic information from free text annotations by extracting entities, symptoms, and solutions into standardized formats. This allows the original free text to remain for technician convenience while generated structured copies enable automated knowledge sharing and case-based reasoning across the system.
Solution Approach 2:
The patent introduces an intermediary processing layer that transforms unstructured free text annotations into structured diagnostic knowledge. This intermediary system uses natural language processing and entity extraction to bridge between the technician-friendly free text format and the system's need for structured, shareable information.
2Measurement precision
If extensive manual parameterization is required, then measurement precision of diagnostic features is improved, but device complexity and time consumption worsen
Solution Approach 1:
The system performs self-service by automatically learning diagnostic feature weights and parameters from historical data without requiring extensive manual configuration. The machine learning models autonomously identify relevant features and their importance, eliminating the need for manual parameter tuning while maintaining high diagnostic precision.
Solution Approach 2:
The patent transforms fixed manual parameters into dynamic learned parameters. Instead of requiring experts to manually set diagnostic thresholds and feature weights, the system automatically adjusts these parameters based on patterns learned from historical diagnostic data, adapting to different equipment and failure modes.
3Productivity
If sensor data and textual information are processed separately, then processing speed is improved, but diagnostic accuracy deteriorates
Solution Approach 1:
The system merges sensor data processing with textual information analysis into a unified diagnostic framework. Both data types are processed through integrated machine learning models that consider their interactions, enabling the system to capture relationships between sensor readings and textual symptoms that would be missed in separate processing approaches.
4Loss of time
If automated case recommendation systems are implemented, then troubleshooting time is reduced, but system complexity increases
Solution Approach 1:
The system creates simplified copies of complex diagnostic cases by extracting key features and similarities, then uses these copied representations for rapid case matching and recommendation. This allows the system to provide automated case recommendations without requiring full replication of complex diagnostic reasoning processes.
Data Source
AI summary
A method and an apparatus for optimizing diagnostics of rotating equipment is provided. The apparatus includes a device for providing status information about status of the rotating equipment over a series of time windows whereby status can be derived from sensor features of at least one available sensor taking measurements during a predefinable time period, a device for using deep learning which combines provided historic sensor information with sequence of events data indicating warnings and/or alerts of the rotating equipment, whereby status information is supplemented with via deep learning predicted probabilities whether a warning and/or an alert has occurred within a time window, device for providing an amount of textual diagnostic knowledge cases, device for extracting semantic information on text features from the textual diagnostic knowledge cases, and device for combining status information and semantic information into a unified representation enabling optimization of the diagnostics.