三维弹塑性有限形变边界元法模拟轧制对界面摩擦的处理

利用金属与轧辊相对流动对摩擦的影响，建立了新的适合于边界元法求解的摩擦模型，该模型在计算中不必对摩擦的方向以及中性点的位置作任何假设，简单实用，有效地解决了接触摩擦难处理的问题，计算结果证明该模型可靠性强。     

摩擦片打滑过程中的热传导模型

摩擦片打滑时，摩擦偶的实际接触是在摩擦表面的微凸体之间进行的，摩擦热的产生具有微区瞬间的特点，从而引起摩擦面局部温度过高，根据摩擦片打滑的特点和导热理论，建立了接触点处的热传导模型，依该模型分析了接触区域温度分布的特点，得知温度沿材料厚度方向呈线性分布，接触点瞬时温度与单位面积摩滑功率和打滑时间成正比，通过验证，用该模型计算所得的结果与文献所得结论相符，并且该模型更能反映摩擦表面温度的实际情况。     

采用混合摩擦模型预报冷轧薄板轧制力

作采用包括粘着摩擦和滑动摩擦在内的混合摩擦模型,研究了冷轧薄板轧制力预报问题,建立了轧制力预报新模型。用新模型对实际生产数据进行了计算,并将理论预报值与实测值和ＭＤＳｔｏｎｅ公式计算结果进行对比,验证了新模型的正确性。     

计算预应力筋摩擦损失和锚固损失的等效荷载法

本文提出计算预应力筋摩擦损失和锚固损失的等效荷载法。用该法计算，不需计算张拉端与计算截面切线的夹角以及曲线筋的曲率半径，只需计算出预应力筋对应的等效荷载，就能简便地计算摩擦应力损失和锚固损失。     

基于数据挖掘的冷轧轧制力优化方法研究

针对广泛应用的Bland-Ford-Hill冷轧轧制力工艺模型,通过挖掘现场实际数据隐含的规律,对其变形抗力和摩擦因数的模型参数进行优化,以提高轧制力计算精度。首先,推导由轧制力计算变形抗力和摩擦因数的逆计算算法,采用L-M非线性多项式回归方法对变形抗力和摩擦因数的模型参数进行优化回归计算,建立轧制力优化算法;然后,根据现场海量的实际数据,采用数据挖掘的方法,使用上述优化方法计算更加符合现场实际的变形抗力和摩擦因数的模型参数。优化结果在线运行后,轧制力精度明显提高。     

冷轧控制模型中摩擦因数模型的开发与应用

摩擦因数模型是冷轧的核心工艺模型之一,目前对其进行的大量研究多是在试验环境下进行的。基于经典摩擦因数模型的工艺原理,综合考虑了轧辊粗糙度、轧制速度等关键因素的影响,结合生产控制的实际情况及控制模型的整体设计,提出了适用的摩擦因数计算模型,进而简要介绍了摩擦因数模型关键参数的调试方式,并提出了一套间接评价模型精度的方法。通过现场测试和长期实际应用证明了模型的控制效果优异。     

惯性摩擦焊过程温度场及变形场的数值模拟

惯性摩擦焊接是一种广泛应用的高效、可靠、节能及环保的固态焊接方法.该方法在航天航空领域已广泛应用.本文分析了惯性摩擦焊接过程高度非线性的特点,在此基础上建立了一个基于二维轴对称的惯性摩擦焊接过程有限元热力耦合计算模型.计算区域划分成四节点等参单元,热物性参数和力学性能参数随温度变化,同时考虑了焊接热流随时间的变化、表面对流和辐射换热以及飞边的影响.利用有限元分析软件MSC Marc对同质材料GH4169-GH4169惯性摩擦焊接接头区的瞬态温度场及应力-应变场的分布进行了计算分析.使用热电偶点对焊接接头部分若干点的温度进行了实时测量.结果表明温度场的计算结果与实测值符合较好.另外,还讨论了工艺参数对摩擦焊接接头区应力分布以及飞边形状的影响.     

铸拉中凝固壳破裂极限的计算

从结晶器内壁表面粗糙度的假设出发，提出了处理结晶器与铸件间间隙接触的方法，得到了结晶器与铸件间的间隙值及摩擦应力值，同时分析了引起凝固壳破裂的原因，建立了判断凝固壳破裂极限条件的模型，并利用此模型对铸拉工艺的极限速度进行了计算，结果表明，计算值与实测值比较接近。     

2030板带冷连轧系统在线控制模型接触摩擦应力分布

采用弹塑性有限元法对带钢冷轧过程中接触区摩擦应力进行了模拟，分析了各种因素对摩擦应力分布的影响。利用数值方法对各影响因素对摩擦应力分布的影响进行参数拟合，从而建立各影响因素和摩擦应力分布间的特定关系，并对计算轧制力数学模型中的摩擦应力分布进行修正，使计算结果更为精确。     

宝钢4号高炉喷煤系统的压降模型

以附加压降法为基础,对整套宝钢4号高炉喷煤系统的压降作了模型计算.分别对水平管、上升管、以及不同曲率半径的弯管等建立了相应的固相摩擦因数经验式,得到了一整套计算喷吹管路压降的模型.与生产装置的运行结果相比,该模型的预测精度较好.这一模型的建立不仅对高炉喷煤系统的设计计算提供了有效的参考,而且还对现有的宝钢4号高炉喷煤系统的节能降耗给予了一定的理论支持.     

Friction Estimation and Roll Force Prediction during Hot Strip Rolling

A mathematical model of friction coefficient was proposed for the roll force calculation of hot-rolled strips. The online numerical solving method of the roll force calculation formula based on the proposed friction model was developed and illustrated by the practical calculation case.Then,the friction coefficient during hot strip rolling was estimated from the measured roll force by force model inversion.And then,the expression of friction model was pro-posed by analyzing the calculation process of stress state coefficient,and the model parameters were determined by the shared parameter multi-model nonlinear optimization method.Finally,the industrial experiments demonstrated the feasibility and effectiveness of the related models.The accuracy of the new roll force model based on the built friction model was much higher than that of the traditional Sims model,and it could be applied in the online hot roll-ing process control.     

Simulation of Strip Rolling Using Elastoplastic Contact BEM With Friction

With rollers as elastic bodies and workpieces as elastoplastic bodies,the rolling problem can be viewed as a friction elastic-plastic contact problem.With fewer assumptions in the simulation of strip-rolling process,a boundary element method（BEM）for two-dimensional elastoplastic finite strain and finite deformation analysis of contact problems with friction was presented.All the equations for contact problems,which include multi-nonlinearities,were obtained.Incremental and iterative procedures were used to find contact pressure and friction stress.Moreover,initial strain rate algorithm and work-hardening material behavior can be assumed in the plastic analysis.Several examples were presented,and the results of contact pressure and friction stress were in excellent agreement with those of analysis.     

Friction Model for Strip Rolling

To improve the accuracy and speed of finite element calculation,a new friction model was proposed,in which a layer of thin friction element between the roller and workpiece was introduced to consider the contact friction.This model was used in the three-dimensional elastic-plastic finite element model of strip rolling process.Taking the geometrical size and physical characteristics of intermediate medium between roller and workpiece into consideration,the stiffness of the friction element was corrected by two factors,i.e.,the thickness of friction element and β factor.Without any assumption about the value and direction of friction force and neutral point position,the rolling process was simulated by using this model.The results obtained by this model were compared with the experimental results and the feasibility of the model was verified.     

Neural Network Friction Model for Cold Strip Rolling

A neural network friction model with very good prediction accuracy was developed on the basis of industrial data. Simulative calculation indicated that the accuracy of rolling force and motor power calculation can be improved using the calculated friction coefficient. It was found that reduction ratio and deformation resistance of strip has more effects than other parameters, and the effects of most parameters are affected by rolling speed.     

Analytical Solution of the Burgers Equation for Simulating Translatory Waves in Conveyance Channels

A Burgers equation model (BEM) for simulating translatory waves in conveyance channels is extracted from the Saint-Venant equations for small perturbations in initial uniform flow. The present study improves upon the previous model and presents analytical solutions for the simulation of translatory waves occurring in conveyance channels. The BEM is reduced to the linear diffusion equation using the Cole–Hopf transformation and then solved by means of the Green’s function assuming an infinite domain. The simulation studies performed show that the BEM results are comparable to those of the Saint-Venant equations for small perturbations in the initial uniform flow conditions and for Froude numbers within the subcritical region. The BEM could be useful for flood routing and for simulating release of water from a reservoir into a conveyance channel when the flow perturbation is small.     

冷连轧高速轧制过程中摩擦因数机理模型的研究

在现场试验与理论分析的基础上,针对现有摩擦模型对冷连轧高速轧制过程中摩擦因数的预报精度不高、通用性不强等问题,首先对润滑油膜厚度计算模型与摩擦因数返算方法进行了推导,然后定量研究了辊缝中润滑油膜厚度与摩擦因数之间的关系,最终建立起一套适合冷连轧高速轧制的摩擦因数机理模型,并将其应用于宝钢冷轧薄板厂1220五机架冷连轧机与宝钢冷轧厂2030五机架冷连轧机的生产实践.应用结果表明:提高了轧制的稳定性与带材的控制精度,减小了相关模型的自学习与自适应系数波动,取得了良好的使用效果.     

提高轧制力设定精度的一种离线优化方法

轧制力参数设定是轧机设定模型的核心参数之一,对决定成品卷的厚度精度及板型质量至关重要。京唐酸轧二级模型计算轧制力时采用Bland-Ford-Hill公式,经分析确定影响轧制力计算精度的参数主要为变形抗力和摩擦力因数。因此提出了一种利用特定钢种的初始历史实际轧制数据离线调整其变形抗力和摩擦力因数的方法,从而提高本钢种在线轧制力模型的设定精度。实际应用表明,使用通过此方法优化后的变形抗力和摩擦力参数计算轧制力,轧制力设定模型的精度得到了明显提高,能够满足在线控制需求。     

基于BP神经网络的机器人波动摩擦力矩修正方法

针对机器人谐波减速器关节在转动过程中存在的波动摩擦力矩, 提出一种基于傅里叶级数函数和BP神经网络的建模方法, 并完善机器人的动力学模型, 修正了因波动摩擦力矩带来的关节力矩计算误差. 通过研究谐波减速器关节的波动摩擦力矩在不同影响因素下的变化特性, 采用傅里叶级数与BP神经网络结合的方法对波动摩擦力矩进行建模. 通过添加傅里叶级数函数作为BP神经网络的辅助输入, 克服了力矩误差曲线因存在高频周期性波动而难以拟合的困难. 在离线环境下训练神经网络, 完成对关节波动摩擦力矩的建模, 进而完善机器人的动力学模型和修正关节中存在的波动摩擦力矩. 验证实验表明, 使用完善后的动力学模型可以有效计算谐波减速器关节的波动摩擦力矩, 并使修正后的力矩误差维持在[-0.5, 0.5] N·m的范围之内, 方差为0.1659 N2·m2, 是修正前的24.23%.      

宽带钢冷轧机颤振控制的数值模拟

本文以较为准确的轧机振动状态下的轧制压力模型为基础，对宽带钢冷轧机干摩擦阻尼减振进行了数值模拟的研究。研究结果与实际观测值相等。     

5052铝合金热粗轧过程温度模型的研究

根据传热学规律,分析了流换热、辐射散热、接触传热,变形热和摩擦热等主要因素对5052铝合金热粗轧温度变化的影响;得出了相应的温度模型。结果表明,温度预测误差不超过1℃,用该温度模型对轧制过程进行仿真计算可以较准确的预测实际轧制温度。