Papers by Keyword: Hot Rolling

Abstract: A full Profile Contour and Flatness Control (PCFC) model prototype has been developed for hot finishing mills. This model prototype accounts for several physical based sub-models calculating the different contributions to the roll gap profile and allows for offline predictions in both preset and recalculation modes. To evaluate the PCFC model developed, an exhaustive comparison analysis between its calculations, the ones coming from the plant model and measures at the finishing mill exit has been carried out. An industrial mill database composed of different rolling campaign types was applied for this purpose and both (i) strip crown and flatness indicators as well as (ii) full strip profiles results have been used for the comparisons. Encouraging results were obtained from this performance assessment since the PFCF model developed leads to similar behavior compared to the existing plant’s model (from an industrial supplier). As a result, the PCFC model developed shows high potential for online implementation in hot strip mills.

Abstract: Kolmogorov-Arnold networks (KANs) have emerged as a promising counterpart to multi-layer perceptrons (MLPs) which offer a more interpretable functionality for different machine learning(ML) applications. Their main difference lies in the definition of KAN layers, using learnable activa-tion functions, which has made these networks optimal for physics-based applications. In this work,we focus on analyzing the performance of KANs in capturing the physics of the hot rolling process,which is an integral part of steel manufacturing industry. Initially, we introduce non-dimensional pa-rameters to encapsulate geometrical factors in the process. We perform space-filling sampling in thespace spanned by these parameters. The sampled points yield the necessary parameters for the finite el-ement (FE) simulations, forming the ground truth (GT) data for the network. A closed-form analyticalmodel for spread is considered from previous studies in the literature, and its predictive performanceis assessed against the FE results. In defining the input space for the network, different alternatives arecompared and it was seen that input space containing the non-dimensional features and the predictionsof the analytical model reduced overfitting and better generalization. The effect of KAN hyperparam-eters are evaluated, and the network with tuned parameters demonstrate optimal performance on thetest set. Lastly, after applying symbolification for this network, a closed-form expression is obtainedthat captures the discrepancy between the analytical model and the GT results, and its performance istested against test set data.

Abstract: This work investigates the use of symbolic regression (SR) to address the trade-off between predictive accuracy and computational efficiency in modeling physical phenomena by constructing compact, closed-form expressions directly from data. In this study, SR is applied to develop fast and accurate models for predicting lateral spread in the hot rolling of steel slabs. The SR models are trained on high-fidelity finite element simulation data and evaluated against established analytical models. Model selection is guided by a parsimony-based optimization strategy that balances predictive accuracy and expression complexity. The results show that the SR-derived formulations achieve lower prediction errors with reduced complexity compared to traditional analytical models. Moreover, SR maintains strong predictive performance even when trained on limited datasets, demonstrating its robustness. Overall, the findings of this work highlight the suitability of symbolic regression for computationally efficient and accurate modeling of complex physical phenomena.

Abstract: Accurate prediction of quality-relevant material parameters, such as thickness and grain size, ensures product quality in hot forming processes. This task becomes especially complex in hot rolling, where the sequential and time-dependent nature of the process results in pass schedules with varying numbers of passes and grain size evolution that depends on the deformation history. To address this complexity, this study aims to develop and train deep learning models based on Long Short-Term Memory (LSTM) networks, which are well-suited for modelling sequential data. As input features for the model, real-time process parameters such as rolling force and rolling temperature are used, which can be captured by sensors during operation. Simulation data for both material and process parameters are acquired using Simulation as a Service (SaaS) through the Fast Rolling Model (FRM) called Rolling Calculation Tool (RoCaT), focusing on steel grade S355. The performance of the LSTM model is evaluated by analysing loss curves over training epochs and comparing predicted values to reference data. The maximum relative percentage error for thickness between the LSTM predicted value and the RoCaT value is 18.818% and 16.56259%, respectively, for pass schedules of 15 and 17 passes, with a starting thickness of 205mm. The percentage of relative error values for grain size is more pronounced during the initial passes as compared to later passes for both pass schedules. The statistical validation is performed on the denormalized data using metrics such as Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and R-Squared (R²), which demonstrate the model’s ability to predict key material properties in the hot rolling process reliably. The RMSE, MAE, and R² values for thickness are obtained as 45.2175mm, 18.7555mm, and 0.8115, respectively. For grain size, the corresponding values are 25.8287µm, 13.4319µm, and 0.9192.

Abstract: Strip casting of Al–25%Si was tested using an unequal-diameter twin-roll caster equipped with steel rolls. To increase the cooling ability, the shell thickness was decreased to 6 mm. The solidification length was 180 mm to solidify Al–25%Si, which has large latent heat. 4% Mg was added to Al–25%Si to prevent the strip sticking to the roll without using parting material. A grooved roll was adopted for the upper roll to prevent solidification shrinkage and strip cracking during rolling. Al–25%Si–4%Mg strip could be cast at a roll speed of 10 m/min by these enhancements. The as-cast strip could be hot rolled down to 1 mm. A cross section of the strip was investigated using optical microscopy, and the primary Si and eutectic Si were found to be very fine.

Abstract: An important aspect of New Product Development (NPD) is the determination of the expected product dimensional window, describing the maximum strip width, that can be produced for a given strip thickness. The estimation of the product dimensional window is used to safely execute first rolling trials. In addition, one can verify in advance whether customer geometry specifications of the final strip can be reached. For this purpose, offline simulation tools are used for hot rolling as well as cold rolling. An accurate prediction of the deformation resistance and interstand softening behaviour of the new steel grade is key in the determination of the dimensional window. Preferably, the deformation resistance model is validated with experimental data, for example from tensile tests or laboratory mini-mills. Rolling simulations are performed, using prescribed process conditions with respect to for example load distributions, temperature and rolling speed requirements. The dimensional windows of respectively the hot strip mill and the cold strip mill are merged, resulting in a final product dimensional window, indicating the maximum strip width at a final, customer specified, strip thickness.

Abstract: Carbide free bainitic microstructures of steels in hot rolled condition have high potential for automotive and structural applications, where both high elongation and toughness at a high strength level are needed. However, achieving a combination of these properties remains a challenge due to difficulties in ensuring a high stability of retained austenite while maintaining industrial processability. Therefore, an attempt has been made in this work to achieve combined high toughness and high elongation in hot rolled carbide free bainitic steels.

Abstract: To attain the aim of weight reduction and safety improvement of vehicles, some high strength steel sheets have been developed and investigated. TRIP-aided steel sheets with transformation-induced plasticity (TRIP) of the retained austenite have high strength and ductility, and excellent hydrogen embrittlement resistance. In previous study, as high strength TRIP-aided steel for forging parts, the volume fraction of retained austenite in the TRIP-aided steel could be increased by hot forging with austempering. Similarly, our research group reported that the thermomechanical process of hot rolling following by austempering could also increase the amount of retained austenite in the TRIP-aided steel sheet. The tensile properties and formabilities of TRIP-aided steel sheet subjected to the thermomechanical rolling just before austempering possess obvious advantages compared with those of TRIP-aided steel sheet without thermomechanical rolling process (with only austempering). These excellent mechanical properties may be caused by the finely dispersed retained austenite and refined bainitic ferrite and/or martensite brock by thermomechanical rolling process.

Abstract: The effect of Ti/Al ratio (mass%) on the evolution of the microstructures after casting and hot rolling of Ca treated 441 dual stabilized ferritic stainless steel (FSS) was investigated in order to understand its effects on grain refinement mechanism. Industrially cast and lab simulated hot rolled samples, were subjected to similar processing conditions but with different Ti/Al ratios of 2.4 and 7.8. The microstructures and inclusions were analysed by the OM, SEM-EDS, SEM-EBSD and AzTecFeature software. The results showed that the steel with higher Ti/Al ratio exhibited finer grains after continuous casting and hot rolling, i.e., the initial finer as-cast structure resulted in finer grains and less substructure after hot rolling. The steel with higher Ti/Al ratio contained more Ti-rich complex inclusions and precipitates (especially TiN), which led to more heterogeneous nucleation of the 𝜹-ferrite and grain refinement during solidification. On the contrary, the steel with low Ti/Al ratio exhibited coarser as-cast grain structure, less recrystallization and higher volume fraction of substructure after hot rolling. Therefore, it was deduced that the Ti/Al ratio is one of the essential parameters to achieving grain refinement in Ca treated 441 FSS during continuous casting.

Abstract: Physical simulation is a well-known tool for cloning industrial rolling conditions and generating large quantities of useful data. Physical simulation allows not only considerable time and resources savings in the development process but also product quality improvement. Additionally, physical simulation enables risk-free experiments and at negligible cost when compared to full-scale trials. In this paper, three physical simulation cases are presented. All cases are applicable to steel hot rolling or post cold rolling: (1) a strip rolling simulation where the roll forces are predicted, (2) a plate rolling simulation aiming improvement in mechanical properties and (3) a continuous annealing line case where the increase in the line speed generated a large raise in profit and productivity.