Three-dimensional simulation of flat rolling through a combined finite element–boundary element approach

This paper presents an innovative approach for analysing three-dimensional flat rolling. The proposed approach is based on a solution resulting from the combination of the finite element method with the boundary element method. The finite element method is used to perform the rigid–plastic numerical modelling of the workpiece allowing the estimation of the roll separating force, rolling torque and contact pressure along the surface of the rolls. The boundary element method is applied for computing the elastic deformation of the rolls. The combination of the two numerical methods is made using the finite element solution of the contact pressure along the surface of the rolls to define the boundary conditions to be applied on the elastic analysis of the rolls. The validity of the proposed approach is discussed by comparing the theoretical predictions with experimental data found in the literature.     

A coupled mathematical model for the rollgap and interstand regions in a hot strip mill

In this paper we consider the process of hot-rolling steel strip.The thickness is reduced by passing the strip through a sequenceof rollers with reductions varying from 40% at the first rollpair to about 10% at the last pair. Between the roll pairs,the strip deflection curve is controlled by a looper roll. Amathematical model of the coupled roll gap and interstand regionis required to assist in the initial setup so as to spread theloads and torques evenly through the roll stands. The modelis of potential value in monitoring the transient behaviourof the rolling process. A review of the various models for eachregion and a coupled complete rolling model is described.     

Multi-step prediction of time series with random missing data

Multi-step prediction is still an open challenge in time series prediction. Moreover, practical observations are often incomplete because of sensor failure or outliers causing missing data. Therefore, it is very important to carry out research on multi-step prediction of time series with random missing data. Based on nonlinear filters and multilayer perceptron artificial neural networks (ANNs), one novel approach for multi-step prediction of time series with random missing data is proposed in the study. With the basis of original nonlinear filters which do not consider the missing data, first we obtain the generalized nonlinear filters by using a sequence of independent Bernoulli random variables to model random interruptions. Then the multi-step prediction model of time series with random missing data, which can be fit for the online training of generalized nonlinear filters, is established by using the ANN’s weights to present the state vector and the ANN’s outputs to present the observation equation. The performance between the original nonlinear filters based ANN model for multi-step prediction of time series with missing data and the generalized nonlinear filters based ANN model for multi-step prediction of time series with missing data is compared. Numerical results have demonstrated that the generalized nonlinear filters based ANN are proportionally superior to the original nonlinear filters based ANN for multi-step prediction of time series with missing data.     

Optimization model of recrystallization hot rolling of titanium-vanadium steels

An optimization model was built based on the data of a pilot-scale (4.5 MN load, 225 KW power capacity) rolling mill to minimize the austenite grain sizes of Ti–V steels, which prevail at the instant of completion of the static recrystallization during hot rolling. A computer program developed for this optimization model was run for the rolling schedule, which is designed according to the complete recrystallization case. An energy optimization model developed previously was applied to different rolling schedules. The grain size optimization results demonstrate the effectiveness of these modelling approaches in terms of final grain size, final plate thickness, measured and computed roll force and torque values for both the design of the thermomechanical schedules which produce plate with fine-grained microstructures, high strength, and notch toughness and the temperature-reduction schedules of conventional controlled rolling.     

The Compaction of Granular Materials by Rolling II. The Effect of Material Compressibility

In an earlier paper a theory for the compaction of an incompressiblegranular material into a cohesive strip by rolling was developed.The present paper deals with the extension of this theory tothe case of a compressible material. The investigation is basedon the model of the plane-strain deformation of an isotropiccontinuum satisfying the Coulomb-Mohr flow criterion in whichthe material density is related to the pressure by a power law.General equations for the determination of the roll pressureare derived and are used to compute the roll force and torquefactors for a particular physical process. The results of thesecomputations are compared with values obtained experimentally.     

Real-time simulation of temperature in hot rolling rolls

Model-based process control is widely used in metal working. Often, simplified real-time capable numerical models replace or support measurement. In this work, we investigate the simulation of temperature in hot rolling rolls based on a 3-D model of a roll which is fast enough to act as soft sensor during operation. The involved discretization of the heat equation in cylindrical space is done via simple finite volumes and propagation in time is done via a forward Euler method. This allows a parallel and efficient implementation for GPGPUs. We validate our code against analytic solutions and measurements, provide a detailed performance analysis, and show simulation results for realistic rolls.     

A Nonlinear Control Concept for Rolling Mills

This contribution is concerned with the nonlinear control of a temper rolling mill based on the exact input/output linearization method. The considered subsystem of the temper rolling plant consists of a four‐high mill stand with a hydraulic adjustment system, and bridle rolls in the entry and the exit section. The demand on the temper rolling process is to establish a certain elongation of the steel strip under the action of the roll force and certain predefined backward and forward strip tensions in order to achieve the desired material characteristics and surface properties of the rolled product. The nonlinear control concept to be presented is characterized by the fact that the torques of the bridle rolls are intended as control inputs for the elongation and master speed control, whereas the hydraulic adjustment system and the main mill drive are used to take effect on the strip tensions. After a brief review of the mathematical model of the mill we will focus on the multi‐input nonlinear tension control.     

The ANN application in FEM modeling of mechanical properties of Al–Si alloy

Artificial neural network (ANN) is a nonlinear dynamic computational system suitable for simulations which are hard to be described by physical models where, rather than relying on a number of predetermined assumptions, data is used to form the model. In order to predict the mechanical properties of A356 including yield stress, ultimate tensile strength and elongation percentage, a relatively new approach that uses artificial neural network and finite element technique is presented which combines mechanical properties data in the form of experimental and simulated solidification conditions. It was observed that predictions of this study are consistent with experimental measurements for A356 alloy. The results of this research were also used for solidification codes of SUT CAST software.     

时间序列模型和神经网络模型在股票预测中的分析

利用MATLAB软件编程建立AR模型、RBF和GRNN神经网络模型,滚动预测上证指数开盘价、最高价、最低价和收盘价与实际价格对比,分析误差.结果表明,3种模型用于股票预测均是可行的,误差很小.AR模型不稳定,对个别预测较准;RBF和GRNN网络训练速度都很快,但GRNN比RBF预测效果好.     

Artificial neural network for discrete model order reduction with substructure preservation

This paper presents a new technique for model order reduction (MOR) that is based on an artificial neural network (ANN) prediction. The ANN-based MOR can be applied for different scale systems with substructure preservation. In the proposed technique, the ANN is implemented for predicting the unknown elements of the reduced order model. Prediction of the ANN architecture is based on minimizing the cost function obtained by the difference between the actual and desired system behaviour. The ANN prediction process is pursued while maintaining the full order substructure in the reduced model. The proposed ANN-based model order reduction method is compared to recently published work on MOR techniques. Simulation results verify the validity of the new MOR technique.     

Thermal analysis of hot strip rolling using finite element and upper bound methods

Thermal analysis of hot rolling process has been studied in this work. A finite element method has been coupled with an upper bound solution assuming, triangular velocity field, to predict temperature field during hot strip rolling operation. To do so, an Upwind Petrov–Galerkin scheme together with isoparametric quadrilateral elements has been employed to solve the steady-state heat transfer equation. A comparison has been made between the published and the model predictions and a good agreement was observed showing the accuracy of the proposed model.     

Simulation of roll grinding system dynamics with rotor equations and speed control

A nonlinear dynamical model for paper machine roll grinding process is investigated through a group of delay differential equations with one constant time-delay. In this model, the time-delay effect is originated from shape error traces on the surface of the roll. The contact interaction of the roll and the grindstone is based on the wear theory, and the lateral deformations of the roll as a simply supported continuous beam element inside a rotational coordinate frame and the rotational rigid body vibration system are considered. The PD-controllers of the roll and the grindstone drives are also included. The numerical simulations for time-history responses provide a view of the stability of this grinding process for the design, analysis and verification of industrial roll grinding measurements in future.     

Improving Prediction of Neural Networks: A Study of Two Financial Prediction Tasks

Neural networks are excellent mapping tools for complex financial data. Their mapping capabilities however do not always result in good generalizability for financial prediction models. Increasing the number of nodes and hidden layers in a neural network model produces better mapping of the data since the number of parameters available to the model increases. This is detrimental to generalizability of the model since the model memorizes idiosyncratic patterns in the data. A neural network model can be expected to be more generalizable if the model architecture is made less complex by using fewer input nodes. In this study we simplify the neural network by eliminating input nodes that have the least contribution to the prediction of a desired outcome. We also provide a theoretical relationship of the sensitivity of output variables to the input variables under certain conditions. This research initiates an effort in identifying methods that would improve the generalizability of neural networks in financial prediction tasks by using mergers and bankruptcy models. The result indicates that incorporating more variables that appear relevant in a model does not necessarily improve prediction performance.     

基于NARX神经网络方法的汽轮机转子关键部位应力预测

汽轮机启动过程中,对转子进行应力预测具有重要意义.为满足在线预测国产某350 MW超临界汽轮机转子关键部位应力的需要,提出了一种基于具有外部输入的非线性自回归(NARX)神经网络的应力预测方法.根据转子实际尺寸建立二维轴对称有限元模型,确定了相应的边界条件,并对有限元计算结果进行验证,得到了转子在冷启动工况下的温度场和应力场,确定了转子的危险点.再根据冷启动规则,构建了288个典型启动工况,并对每个启动工况分别进行有限元计算,获取了每个工况下机组运行参数以及转子危险点处应力的时间历程数据,进而建立了神经网络训练样本库,并使用NARX神经网络对危险点处应力进行预测.此基于NARX神经网络的预测方法可以准确预测出汽轮机转子危险点处的应力变化趋势,其预测值与有限元仿真结果吻合良好,可以满足转子应力在线预测的需求.     

A Prediction Model for Skin Wound Suture Forces With Uncertain Material Parameters北大核心CSCD

To evaluate the forces required for the suture of skin wounds quickly and effectively, the nonlinear finite element method was used to calculate the suture forces for skin wounds with different sizes and material parameters. With the calculated results as samples, the prediction model for skin wound suture forces was constructed by means of the EBF neural network model. Given the uncertain skin material parameters influencing the reliability of numerical results, the Monte-Carlo method was used to analyze the uncertainty propagation of skin material parameters. Finally, the prediction analysis and measuring experiment of wound suture forces were carried out with pig skin specimens to verify the reliability of the method. The results showed that, the suture force increases first and then decreases according to the suture point sequence, and the peak force occurs before the center of the wound. For a 40 mm×10 mm wound, the peak suture force is about 1.7 N, and that for a 40 mm×14 mm wound is about 2.5 N. Influenced by the uncertainty of material parameters, the prediction results of suture forces fluctuate by as much as ±0.6 N. The proposed theoretical prediction model provides an effective solution to the problem of parameter uncertainty propagation for biological soft tissue materials such as skins, and makes an important mechanical reference for robotic surgical suture. © 2023 Editorial Office of Applied Mathematics and Mechanics. All rights reserved.     

飞行事故的一种自适应模糊神经网络预测方法研究

飞行事故预测对于预防飞行事故具有十分重要的意义.首先系统分析了空军飞行事故的主要影响因素,对其中的定性因素进行了量化;然后利用系统分析的成果和历史统计数据建立了空军飞行事故的自适应模糊神经网络预测模型.整个预测过程突破了纯数学模型预测的局限性,实现了预测的定性和定量的结合;由于预测中使用了一种基于高木.关野模糊模型的自适应模糊神经网络,从而使预测模型具有很强的自适应能力,预测结果也比较令人满意.     

三种生成神经网络拓扑结构的方法在股票、商品价格预测中的应用及结果比较

对比了三种不同神经网络模型的生成方式:传统神经网络生成模型,遗传算法训练神经网络模型,以及在第二种方式训练参数的基础上,再使用传统神经网络优化生成模型.论文使用上述三种方法对代表性股票和商品价格进行拟合并预测,通过预测结果准确性和稳定性的比较发现:引入遗传算法后的神经网络在样本内的拟合误差有所降低,而第三种方法在样本外有最低的预测误差和最优稳定性.     

An Eulerian‐Lagrangian Concept for the Numerical Simulation of Flat Steady‐State Hot Rolling Processes

For the finite element analysis of stationary flat hot rolling processes, a new and efficient mathematical model was developed. The method is based on an intermediary Eulerian‐Lagrangian concept, where an Eulerian coordinate is employed in the rolling direction, while Lagrangian coordinates are used in the direction of the thickness and width of the strip. This approach yields an efficient algorithm, where the time is eliminated as an independent variable in the steady‐state case. Further, the vector of independent field variables consists of a velocity component in Eulerian and of displacement components in Lagrangian directions. Due to this concept, the free surface deformations can be accounted for directly and the problems encountered with pure Eulerian or Lagrangian models now appear with reduced complexity and can thus be tackled more easily. The general formalism was applied to different practical hot rolling situations, ranging from thick slabs to ultra‐thin hot strips. (© 2004 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)