Learning Process Control with Model Switching to Cut Energy Loss

Solution Overview

Problem

Self-optimizing control strategies, such as extremum seeking control, often result in energy loss and equipment wear due to continuous variation of signals to find optimal operating conditions, making it challenging to develop robust process control systems.

Innovation Solution

A system that uses a processing circuit to learn a steady-state relationship between manipulated and output variables, switching from a self-optimizing control strategy to a model-based control strategy based on the learned relationship, minimizing energy consumption and reducing equipment wear.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If self-optimizing control strategy is used to find optimal operating conditions, then system performance is improved, but energy loss and equipment wear increase

Engineering Contradiction:

Improvesystem performanceVSAvoidenergy loss

Core Design Contradiction:

ProductivityVSLoss of energy

Solution Approach 1:

The system performs preliminary learning action during a training period to build a steady-state model of the process. This preliminary action captures the optimal operating characteristics without continuous signal variation, enabling subsequent operation to use the stored model instead of ongoing self-optimizing searches, thereby reducing energy loss while maintaining performance

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system creates a copy of the steady-state relationship between manipulated and output variables during the training period. This copied model is then used for control decisions instead of continuously performing the self-optimizing search, eliminating the need for ongoing signal variations while preserving the optimal operating points

Inventive Principle:

Principle #26Copying

2Productivity

If self-optimizing control strategy is used to find optimal operating conditions, then system performance is improved, but equipment wear increases

Engineering Contradiction:

Improvesystem performanceVSAvoidequipment wear

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The system performs preliminary learning action during a training period to build a steady-state model of the process. This preliminary action captures the optimal operating characteristics without continuous signal variation, enabling subsequent operation to use the stored model instead of ongoing self-optimizing searches, thereby reducing equipment wear while maintaining performance

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system creates a copy of the steady-state relationship between manipulated and output variables during the training period. This copied model is then used for control decisions instead of continuously performing the self-optimizing search, eliminating the need for ongoing signal variations while preserving the optimal operating points

Inventive Principle:

Principle #26Copying

3Productivity

If self-optimizing control strategy continuously varies signals to find optimal conditions, then optimal operating conditions are maintained, but system stability deteriorates

Engineering Contradiction:

Improveoptimal operating conditionsVSAvoidsystem stability

Core Design Contradiction:

ProductivityVSStability of the object's composition

Solution Approach 1:

The system performs self-optimizing control actions periodically during a training period to learn the steady-state relationship, then switches to using the learned model for the remaining operation. This periodic learning followed by stable model-based operation reduces continuous signal variation while maintaining optimal conditions

Inventive Principle:

Principle #19Periodic action

Solution Approach 2:

The system creates a copy of the steady-state relationship between manipulated and output variables during the training period. This copied model is then used for control decisions instead of continuously performing the self-optimizing search, eliminating the need for ongoing signal variations while preserving the optimal operating points

Inventive Principle:

Principle #26Copying

Data Source

PatentUS8909359B2Process control systems and methods having learning features

Publication Date: 2014.12.09 TYCO FIRE & SECURITY GMBH

AI summary

A system for operating a process includes a processing circuit that uses a self-optimizing control strategy to learn a steady-state relationship between an input and an output. The processing circuit is configured to switch from using the self-optimizing control strategy to using a different control strategy that operates based on the learned steady-state relationship.