Application of different CFD commercial software t

2022-06-14
  • Detail

Comparison of different CFD commercial software for bulb tubular turbine

Abstract: at present, the more popular CFD commercial software include cfx5tm, cfx-tascflowtm, fluent TM and star CDTM. For most users, they don't know the difference of calculation results from different commercial software. Generally speaking, CFD calculation results are largely affected by lattice number, lattice geometric description, turbulence model, discretization method and the natural attributes of the problem itself. This paper introduces the joint calculation of four blade bulb tubular turbine including headrace chamber, guide vane and runner by using cfx5tm and star CDTM in the same grid under steady flow state, and discusses the calculation results

key words: bulb tubular turbine; Lattice; Solution method; Efficiency; Head

1 preface

at present, the more popular CFD commercial software in the world include cfx5tm, cfx-tascflowtm, fluent TM and star CDTM. For most users, they don't know the difference of calculation results from different commercial software. Generally speaking, CFD calculation results are largely affected by lattice number, lattice geometric description, turbulence model, discretization method and the natural attributes of the problem itself. Starting with the lattice number, turbulence model and solution method, this paper introduces the joint calculation of four blade bulb tubular turbine including headrace chamber, guide vane and runner by using cfx5tm and star CDTM under steady flow state. In this paper, the same lattice is used in the joint comparative calculation, and the influence of several groups of lattice numbers on the calculation results is compared, so as to determine a set of appropriate calculation lattice. Finally, according to the calculation results, the parameters such as head and hydraulic efficiency of the turbine are calculated, and the numerical results are compared with the experimental results, which are in good agreement

2 Influence of grid number on calculation efficiency and head

it is well known that CFD calculation results are largely affected by grid number. In order to understand this effect more clearly, five lattices with the number of lattices of 257880, 391600, 531200, 615200 and 712000 were created respectively

first create grid files from star CDTM, and then convert from star CDTM to cfx-tascflowtm and cfx5tm. Star CDTM can output patrantm format lattice files (without face definitions). Tascflowtm can directly identify such grid files. Then add face definition in GCI file of tascflowtm. Cfx5tm can recognize GRD, GCI and BCF files from tascflowtm. Through the above operations, the lattice structure and lattice number are exactly the same in different software

the bulb tubular runner with existing test results is numerically simulated according to the actual test conditions. The steady-state k-epsilon turbulence model and standard wall function are used for simulation. In different softwares, the standard k-epsilon turbulence model and the standard wall function are not exactly the same. The maximum numerical residual of computational convergence is 0.0001. The convection term of the governing equation is solved by the first-order upwind difference method. All calculation results and test results in Figure 1 are dimensionless

Fig. 1 Comparison of relative efficiency and experimental force applied by phase servo electromechanical driving ball screw on water head

although the water head and efficiency are different in different software, the overall trend of their change with the number of grids is similar. However, the calculated hydraulic efficiency and head are still very different from the test results. The main reason for this situation may be the first-order upwind difference method. In addition, there are two possible reasons. One reason is that there is no adjustment in the calculation. In the actual design and calculation, if the integral head is greatly different from the given head, the guide vane angle and micro rotating blade angle are usually adjusted according to experience to reduce the head deviation; Another reason is that the number of lattices is not enough and the quality of lattices is not high enough, but it is impossible to increase the number of lattices infinitely, and the improvement of lattice quality is also limited by the actual geometry. The figure shows that when the number of grids reaches 615200, the change of head and efficiency in different software has been very small. Compared with cfx5tm, the calculation results of star CD are more affected by the number of cells

3. Application of first-order difference algorithm

in different software, only by applying the calculation obtained by the same lattice and then adding alternating load to compare the experimental results, can a more reliable comparison conclusion be obtained. Therefore, we use 712000 grids to simulate the bulb tubular runner. Compared with the same calculation conditions, the first-order upwind difference method is used to solve the problem under the steady flow state, and the maximum calculation residual is 0.0001. The turbine has 16 guide vanes and 4 runner blades

3.1 comparison of optimal working conditions

the optimal working condition is: unit speed N11 = n/H0 5 is 458.7 rpm, unit flow Q11 = q/H0 5 is 0.2 m3/s. The GVO at the guide vane mouth is 82.5%, that is, the guide vane angle is 50, the runner blade angle is 20, the water head is 4m, and the maximum efficiency of the runner is close to 94%

3.1.1 comparison of efficiency and head

calculation results of comparison of relative efficiency and relative head are shown in figures 2a and 2b

Fig. 2A comparison of relative head under optimal working condition Fig. 2B comparison of relative efficiency under optimal working condition

the results show that CFX5 has higher efficiency and head than star CD

3.1.2 comparison of other physical quantities

figures 3a to 5B show the pressure field, velocity field, velocity vector and tke (turbulent kinetic energy) of the middle cylindrical section from the inner ring to the outer ring. Figures 6a to 7b show the TKE on the blade surface. Figures 8A to 8b show the blade surface pressure. The graphical results show that the cfx5tm and star CDTM show similar results in terms of pressure field, velocity vector and tke comparison

figure 3A pressure field of star CD intermediate cylinder section Figure 3B pressure field of CFX5 intermediate cylinder section

figure 4A velocity field of star CD intermediate cylinder section figure 4B velocity field of CFX5 intermediate cylinder section

figure 5A velocity vector of star CD intermediate cylinder section figure 5B velocity vector of CFX5 intermediate cylinder section

figure 6A tke of star CD blade working face figure 6B tke of CFX5 blade working face

figure 7a tke of back of star CD blade figure 7b tke of back of CFX5 blade

figure 8 A pressure field of star CD blade working face Fig. 8b pressure field of CFX5 blade working face

3.2 comparison of different operating conditions

the comparison results of relative efficiency and relative water head under different unit speed, different guide vane opening and the same blade angle (i.e. there is no obvious vortex at the runner outlet) are shown in FIG. 10

Fig. 10 comparison of relative efficiency and relative head

the comparison results show that the results of cfx-5tm and star-cdtm are very similar to the overall trend of test results. Compared with star-cdtm, the calculated head of cfx-5tm is higher and the blade circulation is larger, but there is little difference; The calculation efficiency is closer to the test results, but lower than the test results. If the clearance loss and friction loss are considered, the difference in efficiency will be greater. However, the difference is not so great that it is unacceptable. In addition, because the numerical calculation results are relatively comparative results, from this point of view, the current solution algorithms cfx-5tm and star-cdtm can still use 1 The Jinan gold testing instrument has been strictly inspected and qualified for the numerical analysis of bulb tubular turbine. Generally speaking, the numerical efficiency obtained by numerical simulation with CFD software should be higher than the experimental results. Therefore, we will consider changing the solution algorithm for further calculation

4. Application of second-order difference algorithm

the convection term of star-cdtm control equation provides a second-order difference algorithm, which can be carried out by setting the variable Mars. The cfx-5tm convection term provides the high resolution option, which is a hybrid algorithm close to the second-order difference algorithm. The maximum convergence residual is 0.0001. Solution under steady flow state. The comparison of relative efficiency and relative head of calculation results is shown in Figure 11

Fig. 11 comparison of relative efficiency and relative head

the results show that cfx-5tm and star-cdtm can well simulate the test results. Whether cfx-5tm or star-cdtm, the second-order difference algorithm is better than the first-order difference algorithm. Moreover, the numerical efficiency obtained by cfx-5tm is higher than the experimental results, which is more in line with the macro understanding

5 Conclusion

CFD commercial software cfx-5tm and star-cdtm are used to simulate the flow pattern in the moving stationary blade of bulb tubular turbine. After comparison, the following conclusions are obtained:

(1) cfx-5tm and star-cdtm can be applied to numerical simulation of bulb tubular turbine

(2) the optimal numerical efficiency obtained by cfx-5tm using high difference algorithm is higher than the experimental value, which is easy to be understood by engineering designers

(3) grid number and grid quality have a great impact on CFD calculation results, so it is necessary to carry out grid correlation analysis in calculation

(4) in order to improve the accuracy of calculation, it is suggested to adopt high-order discrete scheme and select appropriate turbulence model in numerical calculation

Liu Wanjiang 1, Wu Xidong 1, Zhenming Kezhen 2, shingu Xiansi 2

1. Harbin Institute of electric machinery, Harbin 150040 China 2. Hitachi Corporation, Hitachi 08202 Japan (end)

上海用脱毛膏脱毛会不会越长越粗
广州脂溢性脱发可以治好吗
广州他达拉非片用后感受
四个月的宝宝肚子胀气拉稀

Copyright © 2011 JIN SHI