节点文献

分离式压力耦合算法的研究及用于裂解与水合管式反应器的模拟

A Study of Segregated Pressure-based Algorithm And Simulations of Cracking And Hydration Tubular Reactors

【作者】 刘波

【导师】 辛峰;

【作者基本信息】 天津大学 , 化学工程, 2014, 博士

【摘要】 管式反应器为化学工业中最常用反应器类型之一。反应器内部流体的流动、传热、传质和化学反应等物理过程的研究是该型反应器设计的重要环节。近年来,随着计算机与科学计算的发展,数值模拟已成为开展这类研究强有力的工具之一。由于绝大多数管式反应器中的流体流动可视作不可压缩流,因而管式反应器数值模拟的实质就在于求解反应器内不可压缩流各控制方程。分离式压力耦合算法为求解不可压缩流Navier-Stokes方程最高效、应用最广泛的方法之一。其中,更以压力修正方法(如SIMPLE系列算法)和分步式算法(如Projection算法)最为有名。本文针对这两种算法现存的一些问题展开了研究,以促进方法的完善和实际应用。从面向对象的编程理念出发,本文开发了基于三角形网格SIMPLE算法的计算机程序,并以若干经典算例验证了程序的可靠性,为算法的进一步研究和化工应用提供计算平台。对无压Projection算法法向中间速度边界条件展开了研究,考察其对计算精度的影响。这对该算法理论的完善和实际应用均有显著意义。本文还就加速离散方程迭代求解速度的块修正技术进行了研究,考察其在管式反应器模拟中的数值特性。经研究发现:采用ADI-TDMA算法求解离散方程时辅以块修正技术可大幅提高边界条件向反应器中心的传递速度,从而起到加速迭代收敛的作用;基于管式反应器求解区域的几何特征,轴向块修正不起作用,径向块修正则可加速非线性问题的全局收敛进程,并且这种作用完全通过加速压力修正方程而体现出来。此外,本文采用二维圆柱轴对称坐标下交错网格上的SIMPLE算法分别对环氧乙烷水合反应器和乙烷裂解炉炉管进行了数值模拟。控制方程离散采用有限体积法,为了获得不低于二阶的计算精度,对流项采用QUICK格式,扩散项则采用中心差分格式。根据文献中提供的动力学模型计算了组分输运方程和能量方程源项。湍流的模拟则是采用k-ε模型,并以壁面函数法进行近壁面处理。通过与文献提供的实验、工业数据相比对,证明数值模拟所得结果与实际情况吻合良好。此外,在乙烷裂解炉炉管的模拟过程中发现:块修正技术的引入使压力产生震荡,这就要求状态方程的求解需采取一定策略方能保证迭代不会发散,只有这样才能最终起到加速收敛的作用。

【Abstract】 Tubular reactor is one of the most commonly used industrial chemical reactor types. Study of the physical processes occurring inside the reactor such as the fluid flow, heat transfer, mass transfer and chemical reactions is an important part of the reactor design. In recent years, with the development of computer technics and scientific computing, numerical simulation has become one of the most powerful tools to carry out such studies. Since the flow phenomenon inside the tubular reactor can be considered as incompressible in most of the cases, thus numerical simulation of the tube reactor is essentially solving each of the governing equations of flow.The segregated pressure-based algorithm is one of the most efficient and the most widely used methods for solving the incompressible Navier-Stokes equations. Two specific forms of this algorithm are the most famous which include the pressure correction method(such as SIMPLE algorithm) and fractional-step method(such as Projection algorithm). In this thesis, studies of these two algorithms have been carried out focusing on some of the existing problems to facilitate the methodological improvement and application. From the object-oriented programming concept, this thesis has developed a computer program for the SIMPLE algorithm with unstructured triangle meshes. This program has been validated by several classic benchmark numerical examples and provides a computing platform for further research and chemical applications. In addition, this thesis has studied the boundary conditions of the normal intermediate velocity of the pressure-free projection algorithm and investigated its impacts on the computational precision, which is conductive to the theoretical perfection and practical application.The thesis also studied the block correction technique which usually used to accelerate the iterative solution of the discrete equations and investigated its numerical characteristics in the case of tubular reactor simulation. It can be found from the study that when the ADI-TDMA algorithm is used combing with the block correction technique for solving discrete equations can significantly improve the transfer speed of the boundary conditions to the center of the reactor, and thus plays a role in accelerating the iteration convergence. Moreover, because of the geometrical distinction of the tubular reactor, axial block correction does not work, on the contrary, radial block correction accelerates the process of global convergence of the nonlinear system and this effect is completely reflected by accelerating pressure correction equation.Additionally, the SIMPLE algorithm with variables arranging on staggered grid has been adopted to simulate the non-catalytic hydration of ethylene oxide reactor and ethane cracking furnace tubes under the two-dimensional axisymmetric cylindrical coordinate. The governing equations have been discretized with the finite volume method. In order to obtain at least second-order accuracy, the convection terms and the diffusion terms have been treated with QUICK scheme and central difference scheme respectively. Source terms of the energy equation and component transportation equation have been calculated according to the relevant kinetic models provided by literatures. The k-ε model has been used for simulation of turbulent flow and wall functions have been employed for near-wall treatment. Results produced by numerical simulation have been proved good agreement with experimental and industrial data provided by the literature. In addition, it has been found from the simulation of ethane cracking tube that block correction technique causes pressure fluctuations and eventually leads to iterative divergence. To cope with this problem, some proper strategy should be used for solving the equation of state, so that the iteration process could be accelerated and reach convergence.

  • 【网络出版投稿人】 天津大学
  • 【网络出版年期】2016年 08期
节点文献中: 

本文链接的文献网络图示:

本文的引文网络