【作者】 陈水宣；

【导师】 傅新；

【作者基本信息】 浙江大学 ， 机械电子工程， 2008， 博士

【摘要】 在热连轧生产与控制中,带钢的温度是直接影响产品尺寸精度及组织性能的重要因素之一。高精度的温度控制不仅可以提高轧制稳定性,也可改变成品带钢的晶相组织,细化带钢晶粒,提高产品力学性能、物理性能和加工性能。结合某钢厂热轧生产线的热连轧工艺,对带钢轧制过程和层流冷却过程的传热现象和冶金行为进行了建模和实验研究。建立了热轧带钢精轧混合式温度模型,准确预测了带钢温度的厚度分布,用于实际生产中的终轧温度预测;开发了热轧冷却过程中的温度和相变耦合数值模型,充分考虑了带钢冷却过程中的物理冶金变化行为和温度对热物性参数的影响,模型用于新带钢产品的研发。本文的主要研究工作和创新如下:1)采用有限差分法计算轧制区的温度变化,替换了传统温度设定系统中计算误差较大的轧制温升和接触热传导模型;采用解析式建立了空冷模型和指数形式水冷模型,并对模型参数进行了钢种和厚度的修正;通过对带钢表面温度和平均温度的换算,建立了差分和解析式并存的混合温度模型,在兼顾效率的同时提高了温度模型的预测精度。生产应用中,采用短时、长时自学习方法对换热系数进行在线修正,进一步提高模型的预测精度和稳定性。2)建立以相变潜热为内热源的温度耦合计算模型,并采用变分法对模型中的非稳态热传导方程进行离散化;考虑层流冷却过程中的相变对带钢温度的影响,建立空冷区和水冷区的奥氏体冷却相变动力学模型,并利用迭代法实现温度和相变的非线性耦合计算;耦合模型中考虑了换热系数与带钢运行速度之间的关系,并引入了变热物性参数的计算。对双相高强钢特殊冷却条件下的温度进行模拟计算和试验,结果表明模型精度达到生产工艺要求。3)以某钢厂热轧线的三种碳钢为实验样本,采用实验手段测定不同温度下的热物性参数,研究了带钢的热导率、比热、热膨胀系数和密度与带钢温度的关系,并采用最小二乘法建立了热物性参数的分段函数,用于温度场定量计算,消除了传统模型中对热物性参数进行常值处理所带来的计算误差。4)建立的热轧带钢精轧混合式温度模型通过某钢厂热轧线的终轧温度设定系统进行验证。仿真和实验结果表明,终轧温度预测误差小于7.5℃,能够满足现场生产应用的需要。更多还原

【Abstract】 In the production and control of continuous hot slab rolling, temperature is one of the most important parameters controlling the kinetics of metallurgical transformations and the flow stress of the rolled metal. Temperature setting models and cooling control of steels with high accuracy not only improve the rolling stability but change the metallurgical structure of steel products and refine the crystal grain and finally increase the mechanical characteristics and physical property as well as workability, in terms of ductility, drawability and formability.Combined with rolling technology of the new built hot slap rolling mill at Baosteel, the heat transfer phenomena and metallurgical behaviors occurring during the rolling and laminar cooling process were experimentally studied and modeled. Hybrid temperature models in the finish stands were built to calculate the thermal profile along the strip thickness direction and precisely predict the finishing and coiling temperatures. Then the on-line temperature models had been evaluated and tested in hot strip production. In addition, considering the thermal influence on variation of steel material properties and metallurgical behaviors, numerical models to describe nonlinearly coupled aspects of temperature and phase transformation had been proposed for research and development of new steel products. The main work and innovation are briefly stated as following.1) A differential model was proposed to replace the former thermal model of rolling conduction and an off-line adaption technique based on steel grade and thickness class was adopted to correct the radiation factor of air cooling model. Furthermore, the original water cooling model was replaced by a new exponential model with on-line adjustment. Through the conversion model of surface and average temperature, the enhanced hybrid temperature model presented higher prediction accuracy as well as acceptable response speed. In addition, the short-term and long-term adaption models in exponential form, as well as a neural network adaption system characterized by ’Relay’ initialization learning, were put forward to adjusting the heat transfer coefficients of models.2) The thermal and metallurgical behaviors of the strip occurring during the cooling process in hot strip mill were analysed and coupled mathematical models of temperature and phase transformation had been built. The variational method is utilized for the discretization of the governing transient conduction-convection equation, with heat transfer coefficients adaptively determined by the actual mill data. To consider the thermal effect of phase transformation during cooling, a constitutive equation for describing austenite decomposition kinetics of steel in air and water cooling zones is coupled with the heat transfer model by iterative procedure. In addition, numerical simulation was performed in the special cooling condition of dual phase steel, with results confirming that the setup accuracy of temperature prediction system was satisfactory.3) It should be noted that temperature models relied heavily on thermo-physical parameters such as specific heat, conductivity and density of the metal being rolled. These properties were very limited in the literature and when available they were often quoted at a single temperature such as room temperature. This immediately compromised the accuracy and relevance of theoretical predictions. Thus, three example carbon steels were chosen for experiments and analyses. Experiments are performed on the electro-thermo-mechanical test system and the method of least squares has been used to statistically model the properties as functions of temperature.4) The temperature proposed had been applied in the new built hot strip mill at Baoshan Iron & Steel Co. Ltd. (Baosteel), with results showing that the finishing temperature prediction accuracy of hybrid temperature models in finish stands and the coiling temperature prediction accuracy of finite differential models in laminar cooling zones were 7.5℃and 9.0℃, separately. The comparison among the new temperature models and that of 2050mm and 1580mm hot strip mill at Baosteel as well as other mills in domestic and abroad indicated that the new temperature models had achieved the international advanced level.更多还原