Facility Control Abnormality Diagnosis Across Evaluation and Operation Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In control systems utilizing multiple machine learning models, it is challenging to determine the main cause of abnormalities, as existing technologies lack clear methods to differentiate between the evaluation model and the operation model when an abnormality occurs, leading to difficulties in identifying the root cause of issues.
Innovation Solution
An estimation apparatus and method that acquires an abnormality index from state data during facility operations, estimating whether the evaluation model or the operation model is the main cause of the abnormality, and outputs instructions for retraining accordingly, allowing for autonomous identification and correction of issues.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If multiple machine learning models are used in a control system, then the system's functionality and control capability are improved, but it becomes difficult to determine the main cause of abnormalities
Solution Approach 1:
The patent segments the abnormality diagnosis task by creating separate analysis paths for the evaluation model and operation model. The estimation unit divides the determination process into distinct segments: one for assessing evaluation model abnormalities and another for assessing operation model abnormalities, allowing independent analysis of each model's contribution to the abnormality.
Solution Approach 2:
The patent introduces an estimation unit as an intermediary component between the machine learning models and the abnormality detection system. This estimation unit acts as a mediator that analyzes state data and determines whether abnormalities originate from the evaluation model or operation model, bridging the gap between complex model operations and interpretable diagnostic results.
2Measurement precision
If the evaluation model is retrained without identifying the true cause, then the evaluation accuracy may be improved, but unnecessary retraining of the operation model occurs, wasting time and resources
Solution Approach 1:
The patent performs preliminary identification of the abnormality cause before initiating any retraining process. The estimation unit analyzes state data and determines whether the evaluation model or operation model is at fault before retraining begins, preventing unnecessary retraining of the wrong model and optimizing the use of time and computational resources.
3Productivity
If the operation model is retrained without identifying the true cause, then the control performance may be improved, but unnecessary retraining of the evaluation model occurs, wasting computational resources
Solution Approach 1:
The patent performs preliminary identification of the abnormality cause before initiating any retraining process. The estimation unit analyzes state data and determines whether the evaluation model or operation model is at fault before retraining begins, preventing unnecessary retraining of the wrong model and optimizing the use of time and computational resources.
4Measurement precision
If manual analysis of state data is performed to determine abnormality causes, then accurate diagnosis can be achieved, but the process becomes complex and time-consuming
Solution Approach 1:
The patent implements a self-service diagnostic system where the estimation unit automatically analyzes state data and determines abnormality causes without requiring manual intervention. The system uses the output of the evaluation model as a reference to autonomously identify whether the evaluation model or operation model is faulty, simplifying the diagnostic process while maintaining high accuracy.
Data Source
AI summary
An estimation apparatus is provided, including an abnormality index acquisition unit for acquiring, as an abnormality index, an evaluation index output by an evaluation model according to state data being input when an abnormality occurs in a facility, among the state data indicating a state of the facility when an operation model is used to control a control target, the operation model being generated by reinforcement learning where an output of the evaluation model trained by machine learning to output the evaluation index in accordance with the state of the facility is set as at least a part of a reward, and outputting an action in accordance with the state of the facility; an estimation unit for estimating which of the evaluation model or the operation model is a main cause of the abnormality; and an output unit for executing an output in accordance with a result of the estimate.