Prediction Control Threshold Adaptation for Production Devices
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing prediction control systems face challenges in determining suitable command values for subject devices when operating in unknown conditions, leading to potential safety issues and reduced performance due to overly conservative preset constraint conditions.
Innovation Solution
A control system that acquires learning data sets to construct a prediction model, estimates the numerical value range of command values, and determines a second threshold value to extend the acceptable range, allowing for safer and more effective operation of subject devices by using a second acceptable range as the constraint condition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a preset constraint condition (first threshold value) is used to limit the command value range for safety, then safety of operation is secured, but the acceptable range becomes narrower than necessary and prediction control performance is reduced
Solution Approach 1:
The system dynamically adjusts the constraint condition from a fixed first threshold value to a variable second threshold value based on the predicted future state. The control value determination unit selects between the first and second threshold values depending on whether the predicted state falls within the learning data distribution range, making the constraint adaptive rather than static.
Solution Approach 2:
The system changes the parameter of the constraint condition by introducing a second threshold value that is determined based on the predicted future state and the distribution range of learning data. This parameter change allows the acceptable range to expand when the prediction is reliable, thereby improving control performance while maintaining safety.
2Reliability
If the first threshold value is set conservatively to ensure safety, then safety is improved, but the acceptable range of command values becomes narrower and may reject valid predictions
Solution Approach 1:
The system makes the acceptable range dynamic by switching between a fixed first threshold value and a variable second threshold value. When the predicted future state is within the learning data distribution, the system adapts by using the second threshold value, which provides a more appropriate and less restrictive constraint.
Solution Approach 2:
The system uses feedback from the prediction result to adjust the constraint condition. The control value determination unit evaluates whether the predicted state falls within the learning data distribution range and accordingly selects the appropriate threshold value, creating a feedback loop that adapts the constraint to the current operating conditions.
3Reliability
If a fixed constraint condition is applied to all operating conditions, then safety is maintained, but the system cannot adapt to unknown cases and prediction control cannot be appropriately performed
Solution Approach 1:
The system transitions from a static constraint condition to a dynamic one that can adapt to different operating conditions. The control value determination unit dynamically selects between the first and second threshold values based on the predicted future state, enabling the system to handle both known and unknown cases appropriately.
Solution Approach 2:
The system changes the constraint parameter based on the operating condition by determining the second threshold value from the predicted future state and learning data distribution. This parameter change enables the system to maintain safety for known cases while adapting to unknown cases by using the appropriate threshold value.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A technology for performing prediction control by which performance of a prediction model can be sufficiently exerted. A control system according to an aspect of the present invention estimates a numerical value range within which a command value can fall from a distribution of second data relating to the command value in a learning data set used to construct a prediction model, and in such a manner that a first acceptable range prescribed by a preset first threshold value with respect to a command value for a subject device is extended, decides a second threshold value with respect to the command value for the subject device on the basis of the estimated numerical value range. Further, in an operational phase, on the basis of an output value from the prediction model, the control system decides a command value for the subject device within a second acceptable range prescribed by the decided second threshold value, and controls an operation of the subject device on the basis of the decided command value.