Machine Learning Control Model for Fast Low-Overshoot Response

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional control systems, such as PID control, face challenges in achieving high-speed control with minimal overshoot and efficient operation in complex industrial processes due to limitations in handling nonlinear dynamics and variable conditions.

Innovation Solution

A learning processing apparatus and method that utilizes machine learning to generate a control model, which outputs manipulated variables corresponding to indicated and process variables, enabling the generation of controlling data that maps these variables to optimal manipulated values, thereby facilitating faster and more stable control operations across various systems.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If conventional PID control is used for temperature control, then the control system is simple and easy to implement, but the control speed is slow and overshoot cannot be minimized

Engineering Contradiction:
Improvecontrol speedVSAvoidcontrol system complexity
Core Design Contradiction:
SpeedVSDevice complexity

Solution Approach 1:

The patent replaces conventional PID control algorithms with a neural network-based control model. The neural network learns optimal control strategies from historical data and provides faster response with minimal overshoot, substituting the traditional mechanical control approach with an intelligent learning-based system.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The control model is trained in advance using historical operation data to learn optimal control patterns. This preliminary learning phase enables the system to quickly respond to new situations without requiring complex real-time calculations, achieving fast control speed while maintaining system simplicity.

Inventive Principle:
Principle #10Preliminary action

2Adaptability or versatility

If conventional control methods are used for complex industrial processes, then the control logic is straightforward, but the system cannot effectively handle nonlinear dynamics and variable conditions

Engineering Contradiction:
Improvehandling of nonlinear dynamics and variable conditionsVSAvoidcontrol model complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent transforms the control approach by changing from fixed control parameters to dynamic parameter adaptation. The neural network automatically adjusts control parameters based on learned patterns from historical data, enabling effective handling of nonlinear dynamics and variable conditions without requiring complex manual tuning.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The control model performs self-learning from historical operation data, automatically adapting to nonlinear dynamics and variable conditions without requiring external intervention or complex control logic. The system serves itself by learning optimal control strategies from its own operational history.

Inventive Principle:
Principle #25Self-service

3Productivity

If faster control response is achieved, then production rates increase, but energy consumption and operational waste increase

Engineering Contradiction:
Improveproduction rateVSAvoidenergy consumption and operational waste
Core Design Contradiction:
ProductivityVSLoss of energy

Solution Approach 1:

The patent incorporates feedback mechanisms where the neural network continuously learns from historical operation data including energy consumption patterns. This feedback loop enables the system to optimize control actions for both speed and energy efficiency, achieving fast response while minimizing waste through learned best practices.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The control system dynamically adjusts its behavior based on learned patterns from historical data. Instead of always operating at maximum speed, the neural network determines optimal control actions that balance production rate with energy consumption, achieving productivity improvements without proportional increases in energy use.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP4068012B1Learning processing apparatus, control apparatus, learning processing method, control method, learning program and control program
Publication Date: 2024.06.12 YOKOGAWA ELECTRIC CORP
  • EP4068012B1 patent drawingFigure 1A
  • EP4068012B1 patent drawingFigure 1B
  • EP4068012B1 patent drawingFigure 2A

AI summary

There is provided a learning processing apparatus comprising: a learning processing unit configured to generate a control model that outputs a manipulated variable corresponding to an indicated variable and a process variable of a predetermined system by means of machine learning; a generation unit configured to generate controlling data that indicates a correspondence relation of a combination of the indicated variable and the process variable to the manipulated variable corresponding to the combination by using the control model; and a supply unit configured to supply the controlling data to a predetermined control apparatus.