Adaptive Process Control via Dynamic Model Switching

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing process control systems face complexity and inefficiency due to the need for multiple process models and automatic regeneration of controllers to match changing process conditions, leading to increased operational costs and reduced performance as equipment degrades.

Innovation Solution

A system architecture that integrates automatic and independent process model identification and monitoring within the native control system, using function block plug-ins and model identification modules to gather data, create models, and adapt control strategies without user intervention, enabling continuous adaptive control and intelligent monitoring.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If multiple process models are used to match changing process conditions, then adaptability is improved, but device complexity increases

Engineering Contradiction:
Improveadaptability to changing process conditionsVSAvoidcomplexity of process control system
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The system dynamically selects and switches between multiple process models based on current operating conditions. The controller automatically identifies which model best represents the current process state and uses that model for control calculations, enabling the system to adapt to changing conditions without manual intervention while managing complexity through automated model selection

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system changes parameters by switching between different process models that represent different operating regimes. Each model contains specific parameters optimized for particular process conditions, and the system automatically transitions between models as process conditions change, maintaining optimal control performance across varying operating states

Inventive Principle:
Principle #35Parameter changes

2Adaptability or versatility

If automatic regeneration of controllers is implemented, then adaptability is improved, but device complexity increases

Engineering Contradiction:
Improveability to match current process conditionVSAvoidcomplexity of automatic controller regeneration
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The controller performs automatic regeneration of control routines using embedded process models without requiring external intervention. The system self-updates by automatically selecting appropriate models, regenerating control parameters, and implementing new control strategies, thereby reducing the need for manual controller maintenance and regeneration while improving adaptability

Inventive Principle:
Principle #25Self-service

3Manufacturing precision

If model-based control techniques are used, then control precision is improved, but device complexity increases

Engineering Contradiction:
Improveprecision of control responseVSAvoidcomplexity of model-based control implementation
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The system implements model-based control selectively for specific critical control loops where precision is most needed, rather than applying complex model-based techniques to all loops. This partial implementation achieves improved control precision for key processes while limiting the overall system complexity increase to manageable levels

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP2045673B1Method and apparatus for intelligent control and monitoring in a process control system
Publication Date: 2016.08.17 FISHER ROSEMOUNT SYST INC
  • EP2045673B1 patent drawingFigure 1
  • EP2045673B1 patent drawingFigure 2
  • EP2045673B1 patent drawingFigure 3

AI summary

A controller includes a control module to control operation of a process in response to control data, a plug-in module coupled to the control module as a non-layered, integrated extension thereof, and a model identification engine. The plug-in detects a change in the control data, and a collects the control data and data in connection with a condition of the process in response to the detected change. The model identification engine executes a plurality of model parameter identification cycles. Each cycle includes simulations of the process each having different simulation parameter values and each using the control data as an input, an estimation error calculation for each simulation based on an output of the simulation and based on the operating condition data, and a calculation of a model parameter value based on the estimation errors and simulation parameter values used in the simulation corresponding to each of the estimation errors.