Fluid Temperature PID Tuning Using Physical Constraints and AI

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing PID controller tuning methods rely heavily on experience and are limited by different manufacturer algorithms, leading to complex computational challenges in industrial applications, making automatic tuning difficult.

Innovation Solution

An apparatus and method using an artificial intelligence neural network algorithm to automatically tune a fluid temperature PID controller, with a setter calculating initial gain values based on physical properties and a tuner adjusting these values to meet control targets, incorporating a primary and secondary gain tuning process with limit ranges to ensure stability.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If traditional PID tuning methods (Zigler-Nichols, Cohen-Coon, Dahlin) are used, then tuning can be performed with basic proportional, integral, or differential terms, but the tuning results are limited by manufacturer algorithms and depend heavily on experience, reducing reliability

Engineering Contradiction:

Improvetuning reliabilityVSAvoidmanufacturer algorithm compatibility

Core Design Contradiction:

ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent replaces traditional mechanical/experiential tuning methods with an artificial intelligence neural network algorithm. The neural network automatically determines optimal PID parameters by learning from process data, eliminating dependence on manufacturer-specific algorithms and operator experience while improving tuning reliability across different control systems.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Extent of automation

If higher functions are used in automatic tuning methods, then automatic tuning capability is improved, but computational volume increases and complexity increases, making application to industrial sites difficult

Engineering Contradiction:

Improveautomatic tuning capabilityVSAvoidcomputational complexity

Core Design Contradiction:

Extent of automationVSDevice complexity

Solution Approach 1:

The patent segments the automatic tuning process into distinct functional modules: a setter module that calculates initial PID parameter values, a tuner module that optimizes parameters using neural network, and a constraint condition setting module that manages computational limits. This segmentation enables automatic tuning while controlling computational complexity through modular design.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary calculation of initial PID parameters using physical process models before applying the neural network optimization. This preliminary action reduces the search space for the neural network, decreasing computational volume while maintaining automatic tuning capability and improving convergence speed.

Inventive Principle:

Principle #10Preliminary action

3Manufacturing precision

If experience-based tuning is used, then simplicity is maintained, but tuning quality and precision are limited, reducing manufacturing precision

Engineering Contradiction:

Improvetuning precisionVSAvoidtuning operation simplicity

Core Design Contradiction:

Manufacturing precisionVSEase of operation

Solution Approach 1:

The patent implements a self-service automatic tuning system where the neural network independently optimizes PID parameters without requiring operator expertise. The system automatically collects process data, trains the neural network model, and determines optimal parameters, achieving high tuning precision while maintaining operational simplicity through automated execution.

Inventive Principle:

Principle #25Self-service

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

This approach reduces human working time required for tuning, improves fluid temperature control performance, and allows for flexible control target settings, enhancing PID controller efficiency and precision across various industrial sites.

Implementation Method 1

a tuner configured to tune the initial value of the gain of the controller using an artificial intelligence neural network algorithm according to a control target

Methodology Applied

Scientific EffectArtificial intelligence neural network algorithm:

Implementation Method 2

a property deriver configured to calculate the physical property of the fluid temperature through an energy conservation equation

Methodology Applied

Scientific EffectEnergy conservation equation:

Data Source

PatentUS11372378B2Apparatus and method for automatically tuning fluid temperature PID controller having physical property of process as constraint condition

Publication Date: 2022.06.28 DOOSAN HEAVY IND & CONSTR CO LTD

AI summary

An apparatus for automatically tuning a fluid temperature PID (proportional-integral-differential) controller is provided. The apparatus for automatically tuning a fluid temperature PID controller includes: a setter configured to calculate an initial value of a gain of a controller configured to control a fluid temperature by deriving a physical property of the fluid temperature, and a tuner configured to tune the initial value of the gain of the controller using an artificial intelligence neural network algorithm according to a control target.