Galvanized Steel Strip Coating Weight Prediction for Stable Zinc Control

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing methods for predicting the zinc coating weight of galvanized steel sheets lack sufficient accuracy, leading to inconsistent and potentially costly production.

Innovation Solution

A method involving a coating weight prediction model that utilizes operational parameters from both the annealing and coating sections, including quality property parameters of the steel strip, trained through machine learning, to accurately predict and control the zinc coating weight.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional database-based methods are used to control zinc coating weight, then the control process can be implemented, but the prediction accuracy of zinc coating weight is insufficient

Engineering Contradiction:

Improveprediction accuracy of zinc coating weightVSAvoidcontrol reliability of zinc coating weight

Core Design Contradiction:

Measurement precisionVSReliability

Solution Approach 1:

The patent applies parameter changes by expanding the input parameters from conventional database-based methods to include multiple operational parameters (annealing temperature, coating temperature, line speed, zinc bath composition) and steel strip properties (thickness, width, surface roughness). This comprehensive parameter set feeds into a machine learning model that dynamically adjusts predictions based on the specific combination of parameters, thereby improving both prediction accuracy and control reliability of zinc coating weight.

Inventive Principle:

Principle #35Parameter changes

2Measurement precision

If machine learning models with multiple input parameters are used, then prediction accuracy improves, but model complexity and data processing requirements increase

Engineering Contradiction:

Improvecoating weight prediction accuracyVSAvoidprediction model complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent implements self-service by using historical production data and operational parameters to automatically train and refine the machine learning model without requiring manual intervention for model creation or adjustment. The system continuously learns from past data, automatically optimizing its internal parameters and structures to improve prediction accuracy while managing complexity through automated processes rather than manual model configuration.

Inventive Principle:

Principle #25Self-service

Data Source

PatentEP4685262A1Plating adhesion amount prediction method, hot-dip galvanized steel strip manufacturing method, plating adhesion amount prediction model generation method, and plating adhesion amount prediction device

Publication Date: 2026.01.28 JFE STEEL CORP

Figure 1
Figure 2~3
Figure 4

AI summary

There is provided a coating weight prediction method which can predict the coating weight of a galvanized steel strip with high accuracy. The coating weight prediction method for predicting a coating weight of a galvanized steel strip in a galvanized steel strip production facility including a coating section for dipping a steel strip, which has been annealed in an annealing section, in a galvanizing bath, includes: a parameter acquisition step of acquiring one or more of the operational parameters of the coating section, and one or more of the operational parameters of the annealing section; and a coating weight prediction step of inputting input data, including the operational parameters acquired in the parameter acquisition step, into a coating weight prediction model, and causing the model to output the coating weight.