Predictive Control Weight Tuning for Setup-Change Production

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional predictive control methods using prediction models struggle with low prediction accuracy when applied to small-scale production devices due to frequent setup changes, leading to a high probability of malfunctions such as defective product production.

Innovation Solution

A control device that includes a first acquisition unit for acquiring a desired value of a control variable, a second acquisition unit for acquiring a measured value of the control variable, a prediction unit for calculating a prediction value using a prediction model, a desired correction unit for determining a desired command value by correcting the desired basic value based on the prediction value and a weight, and a weight optimization unit for optimizing the weight to ensure appropriate operation of the device.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Extent of automation

If predictive control using a prediction model is implemented on small-scale production devices with frequent setup changes, then advanced control capabilities are achieved, but prediction accuracy deteriorates due to insufficient data collection

Engineering Contradiction:
Improvepredictive control capabilityVSAvoidprediction accuracy
Core Design Contradiction:
Extent of automationVSMeasurement precision

Solution Approach 1:

The system performs preliminary actions by collecting data during periods when the device operates in a fixed state (same workpiece type, same setup). This data is stored and used to generate prediction models in advance, so that when setup changes occur, the pre-collected data can still be utilized for maintaining prediction accuracy without requiring immediate new data collection

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system changes parameters by dynamically adjusting the data collection strategy based on setup stability. When setup changes are detected, the system modifies which data is collected and how it is used for model generation, allowing the prediction model to adapt to different operational conditions while maintaining accuracy

Inventive Principle:
Principle #35Parameter changes

2Measurement precision

If data collection is performed continuously to improve prediction model accuracy, then prediction accuracy improves, but data quality deteriorates due to biased samples from infrequent setup variations

Engineering Contradiction:
Improveprediction accuracyVSAvoiddata representativeness
Core Design Contradiction:
Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The system performs preliminary data collection during stable operational periods, storing this data for later model generation. This allows comprehensive data accumulation without requiring continuous diverse setup changes, ensuring both data quantity and representativeness

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system dynamically adjusts data collection based on detected setup changes. When changes are detected, it modifies collection strategies to capture new operational characteristics, ensuring the data remains representative of current device conditions while maintaining prediction accuracy

Inventive Principle:
Principle #15Dynamics

3Adaptability or versatility

If the prediction model is updated frequently to adapt to setup changes, then adaptability improves, but system complexity increases due to continuous model retraining

Engineering Contradiction:
Improvemodel adaptabilityVSAvoidcontrol system complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The system updates the prediction model periodically based on detected setup changes rather than continuously. This periodic update approach maintains model adaptability to different workpiece types and setup configurations while significantly reducing computational complexity and processing requirements compared to continuous retraining

Inventive Principle:
Principle #19Periodic action

Data Source

PatentEP3933514B1Control device, control method, and control program
Publication Date: 2025.01.29 OMRON CORP
  • EP3933514B1 patent drawingFigure 1~2
  • EP3933514B1 patent drawingFigure 3
  • EP3933514B1 patent drawingFigure 4A~4B

AI summary

The present invention reduces the probability of malfunction occurrence when performing predictive control of a device being controlled. In this control device of one aspect of the present invention, a prediction model for a control variable is used to calculate a prediction value from a measured value of the control variable, and a desired command value of the control variable is determined by correcting a desired basic value in accordance with the calculated prediction value. The degree of correction is determined on the basis of weight. The control device controls the operation of the device being controlled according to the determined desired command value. The control device assesses whether the device being controlled is operated appropriately on the basis of monitoring data relating to the operation result of the device being controlled, and optimizes the weight of the correction to make appropriate control possible based on the assessment result.