Sensor Monitoring Models for Detecting Abnormality Drift
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing failure sign monitoring methods in plants struggle to accurately detect abnormalities due to the inappropriateness of models generated from time-series data, leading to undetected issues and lack of model validation.
Innovation Solution
An information processing apparatus that acquires time-series data from sensors, generates multiple models corresponding to overlapping periods, and uses a detection unit to identify abnormalities by calculating differences between sensor output values and model predictions, with an exclusion unit to remove outdated models and improve detection sensitivity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a single model is generated from time-series data to monitor failure signs, then the monitoring system is simple to operate, but the detection accuracy decreases when the model becomes outdated or inappropriate
Solution Approach 1:
The patent divides the monitoring system into multiple segments by generating multiple models from different time periods (first period, second period, third period) instead of using a single model. This segmentation allows the system to compare sensor data against multiple models, improving detection accuracy when individual models may become outdated or inappropriate for current conditions.
2Measurement precision
If multiple models are generated from different time periods, then the abnormality detection accuracy is improved, but the system complexity increases
Solution Approach 1:
The patent performs preliminary actions by pre-generating multiple models from historical time-series data during different periods before actual monitoring begins. These pre-generated models are stored and ready for comparison, allowing the system to improve detection accuracy without adding significant operational complexity during real-time monitoring.
3Use of energy by moving object
If a single model is used for monitoring, then the system requires less computational resources, but outdated models cannot be identified or replaced
Solution Approach 1:
The patent implements periodic action by generating models at different time periods (first period, second period, third period) and systematically comparing sensor data against each model in sequence. This periodic model comparison approach allows the system to identify when a model becomes outdated or inappropriate while maintaining reasonable computational efficiency through structured, periodic evaluation rather than continuous reanalysis.
Data Source
AI summary
An information processing apparatus includes an acquisition unit configured to acquire a plurality of time-series data indicating changes in output values of a plurality of sensors from a monitoring target including the plurality of sensors, a model generation unit configured to generate one model indicating a relationship between the plurality of time-series data from each of a plurality of periods in the plurality of time-series data, thereby generating a plurality of models respectively corresponding to the plurality of periods, and a detection unit configured to detect an abnormality of the monitoring target based on the plurality of models and the plurality of time-series data. The plurality of periods can include two periods partially overlapping each other.