1. Introduction
During continuous casting, molten steel is poured into a mold that has a certain internal shape; the slab in the mold half-solidifies during passage through the mold, and is continuously extruded from the lower side of the mold to produce semi-finished products of various shapes [
1,
2]. Hot-charge rolling (HCR) is a variant that can save energy and manpower by charging a high-temperature cast slab from a casting machine directly into a rolling furnace [
3,
4]. HCR requires cast steel that has no surface defects. However, the operating factors to prevent cracks are not easily controlled, so the cast steel frequently has surface defects [
5]. In particular, steels for marine structure, shipbuilding, and pressure vessel are more vulnerable to cracks [
6]. This problem requires inspection of the cast slab, and additional processes such as scarfing and grinding, which decrease the productivity of the continuous-casting process and increase the production cost.
Most commercially-produced steel grades have U or V-type ductility curve, as measured using reduction in area (RA) of the cross-section of the gauge section of a specimen at fracture in a tensile test. At the slab cools, ductility is high at temperatures 1100 ≥
T ≥ 1000 °C, decreases at 800 ≥
T ≥ 700 °C, then increases again at 700 °C >
T. If
TMIN-DUCT at which the ductility of steel is lowest can be accurately predicted, surface cracking of the cast steel can be minimized by controlling
T in the casting machine to be >
TMIN-DUCT in the bending area and <
TMIN-DUCT in the unbending area [
6]. However, some steel grades that have
N-type ductility behavior instead of U or V-type, ductility decreases at low temperatures, so the complexity of
TS control increases.
To minimize the occurrence of cracks on the slab surface, the surface temperature
TS of the slab must not be allowed to enter the low-ductility range during the bending and unbending processes that occur in the continuous caster [
2,
5,
6,
7,
8,
9]. Tensile tests of each steel can identify the
T range in which it is brittle (Brittle range). The tests must be performed several times at various
T, so the time and cost are high [
10,
11]. Too high cooling rate and alloying amount may reduce cast steel plasticity and lead to crack formation [
2,
11,
12]. Alloying also has an important effect on the ductility of cast steel [
10,
13]. Even at the same composition, the brittle range can be changed by the thermal history and by the stress applied to the cast steel [
7,
9,
12,
13]. Therefore, tests to find the brittle range of the steel must be repeated at different conditions including strain rate and cooling rate. AS a result of these complications, tensile testing of various steel types to identify the brittle range is not practical [
6,
11,
14,
15,
16,
17].
An alternative approach is to develop a model that can predict high-temperature ductility without needing a complicated experimental process. The RA prediction is to minimize cracks of the slab by avoiding the temperature section with low ductility at the bending and unbending stage of the continuous casting process. Methods that have been proposed for this purpose include linear regression [
14], multiple linear regression [
15], a back-propagation neural network (NN) [
16], and deep neural network with a Gaussian curve [
6,
11]. Existing studies [
6,
11,
14,
15,
16] have focused on alloy steels in which the ductility behaviors have the general U- or V-shape.
Deep neural network [
6,
11] collected RA data from web-based academic database and used the Gaussian fitting since more than 70% of the collected data had a U-or V-typed RA pattern. The study converted measured RA values into the low temperature limit (LTL), central temperature (CT) and high temperature limit (HTL) by using only steel grade data with U- or V-shaped ductile behavior from the collected database. The neural network model was selected as the best performance than the other three models such as random forest, gaussian process, support vector machine in all three indicators: LTL, CT, and HTL. However, this prediction model had a limitation in predicting the hot ductility of cast steel with low-temperature transformation structure during continuous casting. When the cooling rate is fast or when various alloy components are added, it is difficult to predict hot ductility with the Gaussian fitting. When a hard phase such as bainite or martensite forms in the cast steel during continuous casting, the ductility of the steel decreases in the low-temperature region [
5]. Few studies have tried to predict the ductility behavior of alloy steels in which the ductility has an
N- or
W-shape.
So, this paper presents a method to predict the ductility behaviors of steel grades that have an N-shaped RA pattern, the method uses the random forest (RF), which is a type of machine learning. The model uses N-shaped data fitting with six parameters to identify a T below which ductility decreases in the low-temperature region. This study trains and evaluates the RF model using only the steel grade data with N and W typed ductile behavior from the database of the deep neural network with a Gaussian curve. In the learning process, RF regressor works by building various decision trees and calculating the final average RA value based on different compositions, process conditions, and RA values. To validate performance of the proposed model, representative three machine learning models such as gaussian process, support vector machine, and neural network were selected and compared with RF in terms of prediction accuracy for six parameters. The RF model is also compared with the existing empirical formula to predict bainite start temperature of alloy steel using content of alloying elements. Finally, the effectiveness of the proposed model is verified by comparing RA behavior between the Gaussian model and the N-fitting model for the experiment result of Nb-added steel.
5. Conclusions
This study proposed a method to predict RA for cast steel with low-temperature transformed structure during continuous casting. To simulate the behavior of decreasing RA value in the low-temperature section with the data of composition and process conditions, an N-shaped fitting method was proposed and the RA was predicted by random forest, one of machine learning. First, the N-shaped RA behavior derived from six parameters and the RA behavior observed in the actual experiments were compared and analyzed. By comparing the predicted values and the observed values for collected steel grades, it was confirmed that the RF model can effectively predict various types of N-shaped RA behavior. Second, the prediction performance of N-shaped RA behavior was compared using RF, GRP, SVR, and ANN. The difference between the predicted values and the observed values of six parameters for four models was calculated and evaluated by using RMSE. In all other parameters except p5, the predictive performance of the RF model was the best. Third, the Bs temperature was predicted to minimize cracks in cast steel with low-temperature transformed structure during continuous casting. The RF model predicted accurately the Bs temperature more than 30 degrees compared to the empirical formula. The RF model can be a practical alternative to optimally control the secondary cooling conditions of continuous caster. Finally, the change in ductility behavior according to the cooling rate of Nb-added steel was observed. Except for low-carbon steel with low cracking, it is the most produced or widely used carbon region in the continuous casting process. when an alloy such as Nb is added, the probability of cracking is high. The RA prediction model using N-shaped fitting not only showed similarly the ductility behavior of Nb-added steel, but also clearly checked that the ductility decreases near 700 °C.
The limitations of this study were also discussed. First, the RF model for predicting N-shaped RA curve used the composition and process conditions of 108 steel grades. In the collected RA database, the number of steel grades with N-shaped ductility behavior was about 10 to 15%. Due to the limited number of steels, there is some insufficiency in the robustness and adaptability of the RF model. If data are continuously secured, it can be solved by learning and upgrading the RF model. Second, it is not easy to classify the two types of steel grades by using data fitting due to the high similarity between the N- and the U/V-typed curves. Data collection is limited by time and cost practically. So, predicting the absolute RA value according to temperature instead of data fitting should be considered in future research.