Computational Studies on Single-Wall Rotating Vaneless Diffuser of a Centrifugal Compressor Stage

Abstract

This paper focuses on the computational study on the effect of single wall rotation of vaneless diffuser, namely hub side wall and shroud side wall of the vaneless diffuser. ANSYS CFX software is used to analyze the flow parameters and performance of the single wall rotating vaneless diffuser. The rotating diffusers are examined at five dimensionless mass flow rates using SVD as the base model. Using the same impeller, SVD, HRVD, SRVD, and RVD configurations are simulated and compared. Static pressure and total pressure in stationary frame are measured to understand the diffusion process in the single wall rotating vaneless diffuser. Based on the investigations, HRVD outperforms all other configurations with overall improved performance in terms of pressure rise and flow dynamics.

11.1 Introduction

Centrifugal compressors have three main components: impeller, diffuser, and casing. A centrifugal compressor is a device which uses an impeller to accelerate the fluid radially and compress the fluid. At the downstream, the diffuser further converts the velocity of fluid into pressure energy. Centrifugal compressor efficiency is impacted by diffuser losses and non-uniformity in flow at impeller's outlet. This results in a large loss in total pressure and a decrease in centrifugal compressor performance. Rotating a portion of the stationary vaneless diffuser walls at the impeller speed significantly reduces the shear forces between the diffuser walls and the flow. As a result, the operating range and performance are improved [1]. Dean and Senoo did research on vaneless diffuser-equipped centrifugal compressors. The results illustrated on the flow at diffuser entry region is irregular, unstable, and three-dimensional. This causes secondary flow zones to emerge in the diffuser flow domain and subpar compressor performances [2].

To minimize the energy losses in centrifugal compressors due to diffusion, many radial diffuser designs are proposed. A rotating vanless diffuser (RVD) with shroud extension is a straightforward and practical design solution that can significantly improve centrifugal compressor efficiency and lower energy losses. It has been found from a research study that RVD had the capacity to smooth out the incoming distorted particles and enhance the performance of the centrifugal compressor. However, it has been proved that impellers with extended shrouds will only benefit high-speed stage to a small extent and have a minimal impact on the efficiency of low-specific-speed stages [3]. Recently initial numerical studies on a forced RVD with a centrifugal impeller compressor stage having exit blade angle β = 90° were completed by Seralathan and Roy Chowdhury [4] and Govardhan and Seralathan [3]. Sivamani and Ghosh [5] also studied a backswept impeller with impeller’s disc that extended 40% past the tip diameter of the impeller blade. The investigations showed a greater static pressure rise was enhanced by reduced stagnation pressure losses [6].

The walls of the vaneless area of an RVD rotate independently of the impeller. Hence, the diffuser speed will be much less than the impeller speed [7]. Blade trimming or blade cutback is another way of accomplishing additional static pressure recovery, but it has failed and proved to be not suitable to achieve the desired static pressure.

Studies had shown that a free-rotating vaneless diffuser with a rotational speed more than 0.50 times the speed of the impeller was more effective in terms of diffusion. Thus, it had been demonstrated that free-rotating vaneless diffusers can produce high static pressure with minimal energy consumption in low-speed centrifugal compressors [8]. The working range of a centrifugal compressor with a vaned diffuser is less than that of a vaneless diffuser. Vaneless diffusers are employed in a variety of applications due to their wide working range, ease of manufacture, and simple design.

However, all earlier studies attempted to understand the diameter ratio of the RVD portion that outperforms all other diameter ratios. Computational research by a few authors focused on the performance of the forced RVD compressor stage with low specific speed and pressure ratio demonstrated the existence of an optimum dimensional details of the centrifugal compressor.

Hence, the focus of this study is to make one wall of the RVD rotate while the other wall remains stationary viz., the hub side wall of the diffuser rotates while the shroud side wall is stationary and vice versa. All the comparison in terms of performance and flow parameters of a compressor are made with base stationary vaneless diffuser and rotating vaneless diffusers. The commercial CFD code, ANSYS CFX [3] is used for this study and the simulations are carried out for different mass flow rates (ϕ = 0.025, 0.045, 0.054, 0.1, 0.25).

11.2 Computational Methodology

Table 11.1 shows the parameters of the compressor and diffuser. Geometrical models are generated using ANSYS Design Modeller with the dimensions listed here. The diffuser is made to rotate at equivalent speed of the impeller (RVD) and make one wall of the RVD rotate while the other wall remains stationary viz., the hub side wall of the diffuser rotates while the shroud side wall is stationary (HRVD) and vice versa (SRVD). All the comparisons are made with conventional stationary vaneless diffuser (SVD). Figure 11.1 shows the impeller chosen for the present study with its dimensional details listed in Table 11.1.

Table 11.1 Dimensional information

Fig. 11.1

Centrifugal impeller

The meshing of impeller geometry is done in ANSYS Turbo grid [4] with elements count of 339,420 and total number of nodes are 368,714. Target mesh size method is used to generate fine meshing. Figure 11.2 shows the meshed computations domain based on single passage approach chosen here to minimize the computational resources [5]. To reduce the expenses of processing and time, the current work examines using single passage approach for simulation as shown in Fig. 11.2, by assuming that fluid flow in each fluid flow passage is periodic. Boundary conditions are steady state circumstances with total pressure in stationary frame at inlet, and the outlet it is given mass flow rate. The non-rotating and rotating domains are linked through frozen rotor interface [6]. In the conventional diffuser, the diffuser walls are stationary, and it is specified as counter rotating wall as the whole domain is defined as rotating domain initially. Every wall has a no-slip border condition [7]. Air (at 25 °C) is defined as the fluid, total energy is invoked for heat transfer model, and the boundary conditions at the walls are indicated as smooth and adiabatic. The turbulence model kω-SST is used to compute the 3D RANS equations [3, 8]. Due to time and computational resource constraints, the convergence threshold is set at 10^–4. All simulations are adiabatic and steady-state conditions. Governing equations such as continuity and momentum are used with advection scheme high resolution as solver performs the simulations [9]. The boundary conditions imposed are specified in Fig. 11.3.

Fig. 11.2

Meshed domain

Fig. 11.3

Boundary conditions

11.3 Results and Discussion

The computational analysis is done for five different dimensionless mass flow rates namely, ϕ = 0.025, 0.045, 0.054, 0.1, 0.25. It is found that rise in mass flow rate results in stagnation pressure losses. The plots for total pressure in stationary frame of reference, and static pressure is plotted for SVD, HRVD, SRVD, and RVD (refer Figs. 11.4, 11.5, 11.6, 11.7, 11.8, 11.9, 11.10, 11.11, 11.12 and 11.13). The outcomes for various mass flow rates are discussed here.

Fig. 11.4

Mass flow rates (kg/s) versus pressures (kPa)

Fig. 11.5

Mass flow rates (kg/s) versus total pressures in stationary frame (kPa)

Fig. 11.6

Total pressure coefficients for varied mass flow rates of SVD

Fig. 11.7

Static pressure coefficients for varied mass flow rates of SVD

Fig. 11.8

Total pressure coefficients for varied mass flow rates of HRVD

Fig. 11.9

Static pressure coefficients for varied mass flow rates of HRVD

Fig. 11.10

Total pressure coefficients for varied mass flow rates of SRVD

Fig. 11.11

Static pressure coefficients for varied mass flow rates of SRVD

Fig. 11.12

Total pressure coefficients for varied mass flow rates of RVD

Fig. 11.13

Static pressure coefficient for various mass flow rates of RVD

As can be seen in Fig. 11.4, SVD, HRVD, SRVD, and RVD are displaying a similar pattern. The pressure is declining faster after mass flow 0.10 kg/s, and a steep curve is seen after 0.10 kg/s mass flow. All the rotating vaneless diffuser concepts such as RVD, HRVD and SRVD are outperforming SVD. The difference in pressure is very small at mass flow of 0.25 kg/s. It is also observed that SVD, SRVD, and HRVD have nearly the same pressure at 0.10 kg/s mass flow.

Figure 11.5 depicts the total pressure variations for all cases, and it is same at the beginning. For RVD, HRVD, and SRVD, the total pressure increases to 0.10 kg/s whereas on the contrary, the total pressure of the SVD is decreasing. In all cases, behavior after 0.10 kg/s in SVD has the steepest slope, followed by the SRVD, and HRVD. The change in the total pressure of RVD is very small and thus it can be seen clearly that it has a very small slope.

In Fig. 11.6, it can be observed that the total pressure varies with negligible change with respect to the mass flow rate for SVD. When the mass flow rate increases, there is a small change in the total pressure up to 0.10 kg/s mass flow rate, but after 0.10 kg/s, there is a steep decrease in pressure.

Figure 11.7 shows the variations in static pressure in stationary frame for SVD. The total pressure increases gradually up to 0.054 kg/s mass flow rate and after that there is steep increase. Figure 11.8 shows the total pressure variations for HRVD. A rise in the mass flow rate results in total pressure constantly changing with small difference. The total pressure decreases up to mass flow rate 0.054 kg/s and after that it increased and then steep decreased from 0.10 to 0.25 kg/s mass flow rates.

Figure 11.9 lists the variations of static pressure rise for HRVD. The static pressure is constantly decreasing with small changes with the mass flow rate and the pressure is decreasing with some small change up to mass flow rate 0.10 kg/s after that it is steeply decreased.

Figure 11.10 depicts the variations in total pressure in the stationary frame for SRVD. The total pressure is decreasing constantly with small changes and the total pressure is decreasing up to 0.054 kg/s mass flow rate after that it is increases at 0.10 kg/s mass flow rate and then decreased steeply at 0.25 kg/s mass flow rate.

Figure 11.11 shows the variations of static pressure for SRVD. Static pressure decreases with small changes with the mass flow rate and the pressure decreases gradually up to 0.10 kg/s mass flow rate and then it is steeply decreased. Figure 11.12 shows the variations in total pressure in stationary frame for RVD. The total pressure is same as the previous cases, and it decreases with small changes with respect to mass flow rate. The total pressure decreases from 0.025 to 0.045 kg/s mass flow rates and after that it increased constantly with small changes and then it decreased steeply from mass flow rate 0.10 to 0.25 kg/s. Figure 11.13 depicts the variations of static pressure for RVD. The pressure coefficient decreases constantly with some small changes with respect to mass flow rate and the pressure is decreasing gradually with mass flow rate from 0.10 to 0.25 kg/s mass flow rates.

For the sake of brevity, the static pressure and total pressure in stationary frame contours of HRVD at mass flow rates 0.025, 0.045 and 0.10 kg/s are shown here (refer Figs. 11.14, 11.15 and 11.16). The energy transfer from the impeller and diffuser and simultaneous diffusion happening within the passage are clearly captured in the contours. Static pressure gradually increases throughout the passage and the maximum pressures are observed at the domain outlet.

Fig. 11.14

Contours of HRVD for mass flow rate 0.025 kg/s

Fig. 11.15

Contours of HRVD for mass flow rate 0.045 kg/s

Fig. 11.16

Contours of HRVD with mass flow rate 0.1 kg/s

11.4 Conclusion

On analyzing the SVD, HRVD, SRVD, and RVD configurations, it is observed that increase in mass flow rate resulted in a decrease in total pressure and pressure. RVD is outperforming than SVD, HRVD, and SRVD. It is observed that SVD, SRVD, and HRVD have nearly the same pressure at a 0.10 kg/s mass flow rate. In all cases, the flow behavior after 0.1 kg/s, the SVD has the steepest slope, followed by SRVD, and HRVD. The change in the total pressure of RVD is very small. Static pressure coefficient for various mass flow rates of HRVD is declining with the increment in terms of mass flow rates. The static pressure coefficients for various mass flow rates of HRVD is reducing at a very small rate but reduces with a steep slope when the mass flow is increased from 0.1 to 0.25 kg/s. On comparing all the cases, HRVD gives an optimum performance with respect to rise in static pressure when compared to the base SVD model.