Cross-Directional Process Models Using SVM Mismatch Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Model-based industrial process controllers face challenges in determining whether their models accurately represent true process behavior, especially in cross-directional processes, where changes in conditions can lead to inadequate control, and performing experiments to improve data quality is often undesirable.
Innovation Solution
The use of a support vector machine (SVM) for closed-loop model identification and mismatch detection, which analyzes routine operating data to identify clusters of model parameters and detect deviations, allowing for automatic detection of model-plant mismatches without external perturbations, enabling updates or creation of new models to maintain process control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If experiments are performed to improve data quality for model identification, then model accuracy is improved, but product quality may be disturbed
Solution Approach 1:
The patent performs model identification and validation using routine operating data before actual process changes are made. By pre-assessing model accuracy with existing data, the system avoids the need for disruptive experiments that would disturb product quality, thus resolving the contradiction between improving model accuracy and maintaining product quality
Solution Approach 2:
The system uses the process's own routine operating data to validate and update models, rather than requiring external experiments. The model-based controller continuously monitors process behavior and automatically detects when model-plant mismatch occurs, enabling self-validation without disturbing the process or product quality
2Reliability
If model-based control is used to improve control performance, then control effectiveness is improved, but the ability to detect model-plant mismatch deteriorates
Solution Approach 1:
The patent implements a feedback mechanism where the model-based controller continuously compares predicted process behavior with actual measurements from routine operating data. When deviations exceed thresholds, the system automatically triggers model re-identification, creating a closed-loop system that maintains both effective control and accurate model detection
Solution Approach 2:
The system dynamically adjusts between using the model for control and validating the model based on process conditions. The model identification and validation processes are activated adaptively when needed, allowing the system to maintain control effectiveness while periodically detecting model-plant mismatch without continuous disruption
3Reliability
If routine operating data is used for model identification, then process disturbance is avoided, but model accuracy deteriorates
Solution Approach 1:
The patent transforms routine operating data into valid model identification data by applying signal processing techniques and selecting appropriate data segments. The system identifies and uses periods of informative operation within routine data, effectively changing the parameters of data utilization to extract maximum model information without disturbing the process
Solution Approach 2:
The patent replaces the mechanical approach of performing physical experiments with a computational approach using routine operating data. Advanced algorithms process existing data to achieve model identification, substituting physical experimentation with information processing to maintain process stability while achieving adequate model accuracy
Data Source
Figure 1
Figure 2~3
Figure 4
AI summary
A method includes obtaining (402) operating data associated with operation of a cross-directional industrial process controlled by at least one model-based process controller (106, 204). The method also includes, during a training period (502a, 502b), performing (406) closed-loop model identification with a first portion of the operating data to identify multiple sets of first spatial and temporal models. The method further includes identifying (408) clusters (604) associated with parameter values of the first spatial and temporal models. The method also includes, during a testing period (504a, 504b), performing (410) closed-loop model identification with a second portion of the operating data to identify second spatial and temporal models. The method further includes determining (412) whether at least one parameter value of at least one of the second spatial and temporal models falls outside at least one of the clusters. In addition, the method includes, in response to such a determination (414), detecting that a mismatch exists between actual and modeled behaviors of the industrial process.