Page 274 - Fundamentals of The Finite Element Method for Heat and Fluid Flow
P. 274

SOME EXAMPLES OF FLUID FLOW AND HEAT TRANSFER PROBLEMS
                        266
                        are available in Chapter 7. In this subsection, two important benchmark problems that
                        are commonly employed in testing codes will be discussed. In addition, a very recently
                        proposed benchmark test case will also be considered.
                        Flow in a lid-driven cavity
                        Flow in a lid-driven cavity is one of the most widely used benchmark problems to test steady
                        state incompressible fluid dynamics codes. Our interest will be to present this problem as a
                        benchmark for the steady state solution. The definition of the problem is given in Figure 9.1.
                        The geometry is a simple square enclosure with solid walls on all four sides. All the walls,
                        except for the top one, are fixed. The top wall is assumed to be moving with a given
                        velocity; therefore, the fluid attached to this wall also moves with the same velocity in the
                        direction shown in Figure 9.1. A pressure value of zero is forced at the node in the bottom
                        left-hand corner of the cavity as shown.
                           In order to demonstrate the influence of mesh density on the solution procedure, six
                        different meshes have been selected for this problem. We start with a very coarse mesh,
                        as shown in Figure 9.2(a), and refine uniformly by increasing the number of elements as
                        shown in the fourth mesh (Figure 9.2(d)). The fifth mesh is generated by refining the mesh
                        along the cavity walls and coarsening the mesh at the centre as shown in Figure 9.2(e).
                        The meshes shown in Figures 9.2(a) to (e) are all unstructured in nature. The sixth and
                        final mesh is a structured mesh of 100 × 100 uniform divisions, as shown in Figure 9.2(f).
                        At this point, the readers are reminded that a structured mesh gives better accuracy as
                        compared to an unstructured mesh for the same number of nodes.
                           A Reynolds number of 5000 is selected to demonstrate the influence of mesh refinement.
                        The initial values of the velocities at all inside nodes are taken as u 1 = 1and u 2 = 0. The
                        pressure is assumed to be equal to zero at the beginning of the computation. The semi-
                        implicit form of the CBS scheme (see Chapter 7) was used to calculate the solution in
                        time for all the six meshes. Non-dimensional time step values, ranging between 10 −3  and
                        10 −2 , were employed in the calculations. In order to achieve a steady state solution, the


                                                        u = 1, u = 0
                                                         1
                                                              2



                                                u 2 = 0                  u 2 = 0
                                                u 1 =                    u 1 =

                                                      p = 0


                                                         u = u = 0
                                                             2
                                                          1
                        Figure 9.1 Incompressible isothermal flow in a lid-driven cavity. Geometry and boundary
                        conditions
   269   270   271   272   273   274   275   276   277   278   279