Method for optimizing coating thickness control target of cold-rolling hot-dip aluminum-silicon-zinc liquid unit

Abstract

The invention discloses a method for optimizing a coating thickness control target of a cold rolling and hot dipping aluminum-silicon-zinc liquid unit, which comprises the following steps of: firstly, acquiring a large amount of field historical data, and establishing a linear regression model by using a machine learning method by comparing the measurement deviation of a field thickness gauge and a factory quality inspection center thickness gauge to predict the coating thickness deviation; the field target zinc layer thickness is reasonably set according to the predicted plating layer thickness deviation value result, and air knife parameters are dynamically adjusted according to the prediction result of the model; and finally, the plating thickness control effect is evaluated regularly, and the linear regression model is subjected to model iterative optimization according to new data. By the adoption of the method, accurate control over the thickness of the standard plating layer in the hot-dip aluminum-silicon-zinc liquid production line can be achieved, the product quality requirement is met, meanwhile, zinc consumption is reduced, and the production cost is saved.

Description

Technical Field

[0001] The invention belongs to the technical field of continuous hot-dip aluminum-silicon-zinc liquid plate and strip production, and particularly relates to a method for optimizing the control target of coating thickness of a cold-rolled hot-dip aluminum-silicon-zinc liquid unit. Background Art

[0002] Continuous hot-dip galvanizing lines are a vital component of the modern steel industry, with products widely used in a variety of applications, including construction, automotive, and home appliances. In these lines, the uniformity of the zinc coating thickness is a key quality indicator. During production, the thickness of the zinc coating on the steel strip is primarily controlled by an air knife. Positioned above the zinc pot, the air knife sprays nitrogen through a nozzle slit to control the thickness and uniformity of the zinc coating after galvanizing. With the increasingly promising market outlook for aluminum-silicon products, demand for aluminum-silicon products is increasing. During the production process, the zinc coating thickness measured by on-site thickness gauges often deviates from that measured by factory quality inspection centers. If the zinc coating thickness is too thin on-site, discrepancies may occur when measured by factory quality inspection centers. Excessively thick coatings result in significant plating solution waste and cost losses.

[0003] Chinese patent CN110565039A discloses a method for controlling the zinc layer thickness of a hot-dip galvanizing unit. The method comprises: collecting process parameters of the hot-dip galvanizing unit for calculating a predicted zinc layer thickness, wherein the process parameters include strip speed, air knife height, blade-to-strip distance, air knife pressure, blade-lip gap, and strip thickness; determining current operating condition changes, calculating all possible zinc layer thickness prediction values corresponding to the current operating condition changes based on the process parameters, finding a set of zinc layer thickness prediction values closest to the set zinc layer thickness value, and adjusting and controlling the corresponding air knife parameters based on the process parameters corresponding to the set of zinc layer thickness prediction values. This patent can adjust the air knife parameters based on historical data, but it cannot evaluate the deviation between the online zinc layer thickness and the zinc layer thickness detected by the quality inspection center, resulting in unqualified samples from the quality inspection center. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for optimizing the coating thickness control target of a cold-rolled hot-dip aluminum-silicon-zinc liquid unit, which can achieve precise control of the standard coating thickness in a hot-dip aluminum-silicon-zinc liquid production line.

[0005] In order to achieve the above object, the present invention adopts the following technical solution: a method for optimizing the coating thickness control target of a cold rolling hot-dip aluminum-silicon-zinc liquid unit, characterized in that it includes the following steps: (1) Data collection and deviation analysis: collect historical data on strip thickness, strip width, strip running speed, and coating thickness measured by the on-site thickness gauge during the cold-rolled strip production process; compare the coating thickness measured by the on-site thickness gauge with the result of the thickness gauge at the factory quality inspection center, and record the deviation between the two; these deviation data will be used for subsequent model training and analysis.

[0006] (2) Preprocess the collected data and use the strip thickness, strip width, strip running speed, and the coating thickness measured by the on-site thickness gauge as input variables of the linear regression model; use the deviation between the coating thickness measured by the quality inspection center and the coating thickness measured by the on-site thickness gauge as the output variable of the linear regression model; establish a linear regression model to predict the deviation between the on-site thickness gauge and the thickness gauge of the factory quality inspection center; The linear regression model is expressed as: y = -0.0052 t -0.0041 w +0.092 d +0.12 x -6.778 in, y represents the predicted coating thickness deviation, t is the strip thickness, w For the strip width, d is the strip running speed, x The coating thickness is detected by the on-site thickness gauge, and -6.778 is the constant coefficient parameter of the model.

[0007] (3) Target coating thickness setting: adjust the target coating thickness setting of the on-site thickness gauge according to the deviation predicted by the linear regression model; monitor the coating thickness of the strip in real time during the production process, and dynamically adjust the air knife parameters according to the prediction results of the model.

[0008] (4) Regularly evaluate the effect of coating thickness control and iteratively optimize the linear regression model based on new data.

[0009] Furthermore, the measurement process of the coating thickness measured by the on-site thickness gauge in step (1) is as follows: collecting instantaneous coating thickness data of all measuring points from the left edge to the right edge of the strip when the thickness gauge moves in a Z-shaped trajectory on the strip surface.

[0010] Furthermore, the data preprocessing in step (2) includes: (i) outlier removal; (ii) data smoothing; (iii) Data standardization.

[0011] Furthermore, in step (3), at least one air knife parameter is dynamically adjusted according to the model prediction result, and the air knife parameters include air knife height, distance between upper and lower surfaces of the air knife, stabilizing roller pressure and correction roller position.

[0012] Furthermore, the iterative optimization of the model in step (4) includes: collecting thickness measurement data from the quality inspection center after each batch of production and using it to retrain the linear regression model.

[0013] By adopting the above scheme, the present invention can achieve precise control of the standard coating thickness in the hot-dip aluminum-silicon-zinc liquid production line, improve product quality, reduce rework and waste caused by substandard coating thickness, and thus save a lot of production costs.

[0014] The method of the present invention can achieve accurate control of the coating thickness of the production line and has the following advantages: 1. This paper applies machine learning to the coating thickness control of the hot-dip aluminum-silicon-zinc liquid production line. By using a model to predict the measurement deviation between the on-site thickness gauge and the thickness gauge in the quality inspection center, more accurate coating thickness control is achieved.

[0015] 2. Utilize existing thickness measurement equipment without adding additional hardware investment, and improve control effects through software algorithm optimization.

[0016] 3. By precisely controlling the coating thickness, the waste of zinc liquid is reduced, and the rework and scrap rate caused by unqualified coating thickness are reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The present invention is further described with reference to the accompanying drawings. However, the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative effort.

[0018] Figure 1 This is a model control flow chart in the method of the present invention. DETAILED DESCRIPTION

[0019] The following describes a method for accurately and rapidly correcting the actual coating thickness of hot-dip aluminum-silicon-zinc strip steel when the target coating thickness changes, as described in the present invention, based on specific implementations and accompanying drawings. By comparing the measurement deviations of on-site thickness gauges with those used by factory quality inspection centers, collecting extensive field data, and using machine learning to establish a deviation setting model, the method rationally sets the target zinc coating thickness on-site. This method not only improves the rationality of the control target for the zinc coating on the edges of hot-dip galvanized steel sheets, reduces zinc consumption, and the amount of corrections required for thick edge defects and zinc spalling defects, but also brings significant economic benefits to enterprises.

[0020] like Figure 1As shown, the present invention provides a method for optimizing the coating thickness control target of a cold-rolled hot-dip aluminum-silicon-zinc liquid unit, and controls the coating thickness in the width direction of the current coil steel according to the following steps.

[0021] First, collect historical data on strip thickness, strip width, strip running speed, and coating thickness measured by the on-site thickness gauge during the cold-rolled strip production process; at the same time, compare the coating thickness measured by the on-site thickness gauge with the result of the thickness gauge at the factory quality inspection center, and record the deviation between the two.

[0022] Specifically, the measurement process of the coating thickness measured by the on-site thickness gauge is as follows: the thickness gauge moves in a Z-shaped trajectory on the surface of the strip, recording the instantaneous zinc layer thickness data of all measuring points from the left edge to the right edge of the strip to reflect the change law of the zinc layer thickness in the width direction of the strip; and obtaining the data of the coating thickness measured by the on-site thickness gauge.

[0023] Secondly, the collected data is preprocessed, and the strip thickness, strip width, strip running speed, and the coating thickness measured by the on-site thickness gauge are used as input variables of the linear regression model; the deviation between the coating thickness of the quality inspection center and the coating thickness measured by the on-site thickness gauge is used as the output variable of the linear regression model; the linear regression model is constructed using machine learning methods. y .

[0024] The specific model is: y = -0.0052 t -0.0041 w +0.092 d +0.12 x -6.778; among which, y Indicates the predicted coating thickness deviation (on-site thickness gauge test value - quality inspection center thickness gauge test value), t 、 w 、 d Represent the strip thickness (mm), width (mm) and speed (m / min), x Use on-site thickness gauge to detect coating thickness.

[0025] Input the preprocessed data into the pre-trained linear regression model y The following is a prediction of the current strip thickness of 0.976 mm, width of 887 mm, speed of 100 m / min, and target zinc layer thickness of 30 g / m 2 The actual zinc layer thickness measured by the on-site thickness gauge is 33.5 g / m 2 Calculation shows that the zinc layer thickness deviation is 2.8 g / m 2 This means that the thickness gauge on site now measures 2.8 g / m higher than the one measured by the quality inspection center. 2 .

[0026] Based on the model's predicted zinc coating thickness deviation, the target zinc coating thickness on site was revised to 32.8 g / m². The operator adjusted the air knife process parameters (air knife distance and air knife pressure) based on the predicted deviation and ensured that the corrected thickness gauge target zinc coating thickness was within the range of 32.8-33.8 g / m².

[0027] Finally, after each batch of production is completed, accurate thickness measurement data is collected from the quality inspection center for model validation and optimization. The model is regularly retrained using new production data, model parameters are updated, and the updated model's prediction accuracy is evaluated using a test dataset.

[0028] The present invention calculates the deviation between the zinc layer thickness of the zinc layer detected by the quality inspection center and the zinc layer thickness, and adjusts and controls the air knife process parameters based on this, thereby reducing the deviation between the coating thickness and the coating thickness set value, and can meet the coating thickness control requirements of various working conditions, and improve the coating thickness control accuracy and uniformity of the hot-dip aluminum-silicon unit.

[0029] Although the present invention lists the above preferred embodiments, it should be noted that although technicians in this field can make various changes and modifications, unless such changes and modifications deviate from the scope of the present invention, they should be included in the protection scope of the present invention.

Claims

1. A method for optimizing the coating thickness control target of a cold rolling hot-dip aluminum-silicon-zinc liquid unit, characterized in that: The steps include: (1) Data collection and deviation analysis: collect the thickness, width, speed and coating thickness of cold-rolled strip in the historical production process; compare the coating thickness measured by the on-site thickness gauge with the result of the thickness gauge at the factory quality inspection center, and record the deviation between the two; (2) Preprocess the collected data and use the strip thickness, strip width, strip running speed, and the coating thickness measured by the on-site thickness gauge as input variables of the linear regression model; use the deviation between the coating thickness measured by the quality inspection center and the coating thickness measured by the on-site thickness gauge as the output variable of the linear regression model; establish a linear regression model to predict the deviation between the on-site thickness gauge and the thickness gauge of the factory quality inspection center; The linear regression model is expressed as: y = -0.0052 t -0.0041 w +0.092 d +0.12 x -6.778 in, y represents the predicted coating thickness deviation, t is the strip thickness, w For the strip width, d is the strip running speed, x The coating thickness is detected by the on-site thickness gauge, and -6.778 is the constant coefficient parameter of the model; (3) Target coating thickness setting: adjust the target coating thickness setting of the on-site thickness gauge according to the deviation predicted by the linear regression model; monitor the coating thickness of the strip in real time during the production process, and dynamically adjust the air knife parameters according to the prediction results of the model; (4) Regularly evaluate the effect of coating thickness control and iteratively optimize the linear regression model based on new data.

2. The method for optimizing coating thickness control targets of a cold rolling and hot-dip aluminum-silicon-zinc liquid coating unit according to claim 1, characterized in that: The measurement process of the coating thickness measured by the on-site thickness gauge in step (1) is as follows: collecting instantaneous coating thickness data of all measuring points from the left edge to the right edge of the strip when the thickness gauge moves in a Z-shaped trajectory on the strip surface.

3. The method for optimizing coating thickness control target of a cold rolling hot-dip aluminum-silicon-zinc liquid coating unit according to claim 1, characterized in that: The data preprocessing in step (2) includes: (i) outlier removal; (ii) data smoothing; (iii) Data standardization.

4. The method for optimizing coating thickness control target of a cold rolling hot-dip aluminum-silicon-zinc liquid coating unit according to claim 1, characterized in that: In step (3), at least one air knife parameter is dynamically adjusted according to the model prediction result, and the air knife parameters include air knife height, distance between upper and lower surfaces of the air knife, stabilizing roller pressure and correction roller position.

5. The method for optimizing coating thickness control target of a cold rolling and hot-dip aluminum-silicon-zinc liquid coating unit according to any one of claims 1 to 4, characterized in that: The iterative optimization of the model in step (4) includes: collecting thickness measurement data from the quality inspection center after each batch of production and using it to retrain the linear regression model.

Images