Abnormality Cause Estimation Using Pre-Divided Event Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing diagnostic techniques for large-sized equipment, such as compressors, face challenges in constructing diagnostic models that account for varying device configurations, connection forms, and measurement parameters across different plants, leading to inefficiencies and inaccuracies in abnormality cause estimation.
Innovation Solution
The proposed solution involves an abnormality cause estimation device that inputs measurement values from sensors into an abnormality determination unit, which then uses pre-divided abnormality event models to estimate the cause of abnormalities in equipment. This approach allows for accurate and efficient estimation of abnormality causes by considering the specific configurations and parameters of the equipment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a diagnostic model is constructed for each specific equipment configuration and plant, then the accuracy of abnormality cause estimation is improved, but the device complexity and time required for model construction increase
Solution Approach 1:
The patent segments the diagnostic model into multiple abnormality event models, each corresponding to a specific type of abnormality event. This segmentation allows the system to handle different equipment configurations and abnormality types separately, improving estimation accuracy while avoiding the need to construct a completely new model for each scenario.
Solution Approach 2:
The patent creates a universal abnormality cause estimation device that can handle multiple equipment types and configurations through a single system. The device uses a database of pre-prepared abnormality event models that can be selected and applied based on the specific equipment being diagnosed, eliminating the need to construct separate diagnostic models for each plant or equipment configuration.
2Measurement precision
If machine learning methods are used to correlate measurement parameters with abnormality causes, then the accuracy of cause estimation is improved, but the quantity of data required increases, which is problematic when failure frequency is low
Solution Approach 1:
The patent performs preliminary action by preparing multiple abnormality event models in advance, each corresponding to different types of abnormality events. These models are created before actual diagnosis occurs, allowing the system to handle low-frequency failures effectively. When an abnormality occurs, the system simply selects the appropriate pre-prepared model rather than requiring extensive data collection and model training at the time of failure.
3Reliability
If the diagnostic model is customized for each equipment configuration and connection form, then the reliability of diagnosis is improved, but the time required for model construction and maintenance increases
Solution Approach 1:
The patent implements a universal diagnostic system that maintains high reliability across different equipment configurations by using a database of pre-prepared abnormality event models. The system can select and apply the appropriate model based on the specific equipment type and configuration, eliminating the need to construct and maintain separate models for each plant while preserving diagnostic reliability.
Solution Approach 2:
The patent uses pre-prepared abnormality event models that can be copied and applied to different equipment instances. Instead of constructing unique models for each equipment configuration, the system maintains template models that can be selected and adapted based on the specific equipment being diagnosed, significantly reducing model construction and maintenance time.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present invention estimates the cause of abnormality of equipment easily and with high accuracy. An abnormality cause estimation device (1) has: a measurement value input unit (2) that receives measurement values obtained from sensors installed in equipment; an abnormality determination unit (4) that determines whether or not there is an abnormality in the measurement values; an abnormality cause estimation unit (6) that estimates the cause of abnormality in the equipment by inputting the presence or absence of abnormality in the measurement values into a selected one of abnormality event models having undergone division in advance on the basis of abnormal events assumed to occur in the equipment; and a result output unit (9) that outputs the estimation result of the cause of the abnormality.