Edge Controller Model Transfer for Real-Time Facility Control

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing control systems for facilities face challenges in efficiently performing learning processing for control models due to limited processing resources at the edge, making it difficult to manage and control complex industrial facilities effectively.

Innovation Solution

A control system architecture that utilizes edge computing with controllers near the facility and a central control apparatus for learning processing, allowing the central apparatus to generate and update control models, which are then transmitted and used by edge controllers for facility control, enabling efficient distribution of processing loads and resource utilization.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If learning processing is performed at the edge controller, then control responsiveness is improved, but processing resources are insufficient

Engineering Contradiction:
Improvecontrol responsivenessVSAvoidprocessing resources
Core Design Contradiction:
SpeedVSQuantity of substance

Solution Approach 1:

The control system is segmented into two parts: edge controllers that perform real-time control execution and a cloud-based platform that performs learning processing. This segmentation allows each component to specialize in its strength - the edge controller maintains responsiveness while the cloud platform provides sufficient processing resources for complex learning tasks.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

A communication interface serves as an intermediary between the edge controller and cloud platform, enabling the edge controller to offload learning processing to the cloud while maintaining real-time control capabilities. The intermediary manages data exchange and coordination between the two systems.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Measurement precision

If complex learning models are used, then control accuracy is improved, but processing load increases

Engineering Contradiction:
Improvecontrol accuracyVSAvoidprocessing load
Core Design Contradiction:
Measurement precisionVSPower

Solution Approach 1:

Complex learning model processing is extracted from the edge controller and relocated to the cloud-based platform. This extraction allows the use of computationally intensive models for high accuracy control while keeping the edge controller's processing load manageable for real-time execution.

Inventive Principle:
Principle #2Taking out (Extraction)

3Measurement precision

If more processing power is allocated to learning, then model accuracy is improved, but control responsiveness decreases

Engineering Contradiction:
Improvemodel accuracyVSAvoidcontrol responsiveness
Core Design Contradiction:
Measurement precisionVSSpeed

Solution Approach 1:

The system segments processing tasks by function and time scale: the cloud platform handles offline learning model generation with high computational power, while edge controllers handle online real-time control with lower processing requirements. This segmentation resolves the trade-off between model accuracy and control responsiveness.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3985460B1Controller, control system and method to control a facility and to generate a control model by learning
Publication Date: 2023.12.06 YOKOGAWA ELECTRIC CORP
  • EP3985460B1 patent drawingFigure 1
  • EP3985460B1 patent drawingFigure 2
  • EP3985460B1 patent drawingFigure 3

AI summary

Provided is a control system comprising a control apparatus including a learning processing unit configured to generate a control model by learning, the control model being configured to calculate control data for controlling a facility according to state data detected by at least one sensor configured to measure a state of the facility, and a model transmission unit configured to transmit the generated control model to a controller configured to control the facility; and a controller including a model receiving unit configured to receive the learned control model from the control apparatus, a state receiving unit configured to receive the state data from the at least one sensor, a calculation unit configured to calculate control data for controlling the facility according to the state data, which is a processing target, by using the control model received by the model receiving unit, and a control unit configured to control the facility by using the calculated control data.