Plant Alarm Prediction Using Virtual Operation Patterns
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In industrial plants using petroleum, petrochemistry, or gases, workers face challenges in predicting alarms accurately, leading to potential inefficiencies and safety issues due to subjective decision-making based on experience rather than objective data analysis.
Innovation Solution
An information processing device that predicts alarms using models generated under different conditions for various operation patterns, sets the reliability of these predictions, and performs display control based on the reliability of the prediction results, enabling more informed operational decisions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If workers use subjective experience to decide operation details, then operational flexibility is maintained, but alarm prediction accuracy deteriorates
Solution Approach 1:
The patent creates a virtual plant that copies the actual plant's operational characteristics and uses this virtual model to generate multiple possible operation patterns. By simulating different operational scenarios in the virtual environment, the system can predict alarms without requiring complex real-time analysis of every possible decision path, thus improving prediction accuracy while managing system complexity.
Solution Approach 2:
The system performs preliminary simulation of multiple operation patterns before actual operational decisions are made. By pre-calculating potential alarm scenarios based on virtual plant operations, the system provides workers with advance information about possible outcomes, improving alarm prediction accuracy without adding complexity to the actual operational decision-making process.
2Reliability
If multiple operation patterns are simulated, then alarm prediction accuracy is improved, but information processing time increases
Solution Approach 1:
The patent extracts and prioritizes only the most relevant operation patterns for simulation based on the current plant state and historical data. By selecting and simulating only the most probable or critical operation patterns rather than all possible patterns, the system maintains high prediction accuracy while significantly reducing the time required for information processing.
3Ease of operation
If reliability metrics are provided for each operation pattern, then decision-making quality is improved, but data processing complexity increases
Solution Approach 1:
The patent provides reliability metrics specifically for each operation pattern based on its individual characteristics and the current plant state. By calculating and presenting reliability information locally for each specific operation pattern rather than a single overall metric, the system enables workers to make more informed decisions about specific operational choices while keeping the data processing complexity manageable through targeted analysis.
Data Source
AI summary
An information processing device uses models generated under different conditions, and predicts alarms occurring in the case in which each of a plurality of operation patterns, which is virtually generated in relation to the operation of an actual plant by a worker, is implemented with respect to the actual plant. Then, the information processing device sets the degree of reliability of the prediction result regarding each of the plurality of operation patterns. Subsequently, based on the degree of reliability of the prediction result, the information processing device performs display control with respect to the alarms.