Multi-Plant Abnormality Assessment for Predictive Fault Diagnosis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing abnormality detection and diagnosis systems for plants, such as power generation facilities, are limited in their ability to integrate data from multiple plants, leading to unreliable detection and delayed responses due to resource constraints and reliance on real-time data for predicting future anomalies.
Innovation Solution
A plant assistance assessment system that acquires data from multiple plants via a network, incorporating modules for abnormality sign monitoring, fault diagnosis, progress prediction, risk assessment, and abnormal phenomenon assessment to provide integrated management, improved detection reliability, and predictive capabilities for future anomalies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If abnormality detection is performed using real-time data from a single plant, then the system can detect current abnormalities, but the detection reliability is low due to limited resources
Solution Approach 1:
The patent combines data from multiple plants into a centralized database, merging limited local resources into a larger pooled resource base. This allows the system to leverage data from numerous plants to improve detection reliability while each individual plant continues to operate independently.
Solution Approach 2:
The centralized database serves multiple functions: storing historical data from multiple plants, providing training data for AI models, enabling cross-plant anomaly detection, and supporting predictive analytics. This multi-functional approach maximizes the utility of available data resources.
2Reliability
If a centralized system integrates data from multiple plants, then detection reliability improves, but the system complexity increases
Solution Approach 1:
The patent introduces a centralized database as an intermediary layer between multiple plants and the analysis system. This mediator handles data collection, storage, and preliminary processing, simplifying the architecture by separating data aggregation from complex analysis functions.
Solution Approach 2:
The system is divided into distinct modular components: data collection modules at each plant, a centralized database for storage, and separate AI analysis modules for processing. This segmentation allows each component to be developed and maintained independently, reducing overall system complexity.
3Measurement precision
If real-time monitoring is used, then current abnormalities are detected, but future abnormalities cannot be predicted
Solution Approach 1:
The patent uses AI models to perform preliminary analysis of historical data patterns, identifying early warning signs and predicting potential future abnormalities before they occur. This allows proactive maintenance planning and reduces response time for anticipated issues.
Solution Approach 2:
The system implements feedback loops where predicted abnormalities and actual outcomes are continuously compared, allowing the AI models to learn from past performance and improve future predictions. This feedback mechanism enhances both current detection accuracy and predictive capability.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A plant assistance assessment system 100 that assesses a plurality of nuclear power generation plants 110 includes an abnormality sign monitoring system 121 that monitors an abnormality sign in a plant, based on plant data of the nuclear power generation plants; a fault diagnosis system 122 that performs fault diagnosis on each of devices in the plants, based on abnormality sign detection information; a progress prediction system 123 that simulates progress of the plants, based on fault information; an abnormal phenomenon specification system 124 that specifies a potential abnormal phenomenon in the plants, based on progress prediction information; a risk assessment system 125 that assesses a risk of a fault in the devices in the plants in a probabilistic manner; and an abnormal phenomenon assessment system 126 that assesses likelihood of occurrence of an abnormal phenomenon, based on abnormal phenomenon information and fault probability information.