Research on the effect of various tongue types on large volute pumps

In order to analyze the effect of different tongue type on large volute pump, this paper takes a large volute pump as the research object, and analyzes the calculation results of three different types of volute models. The commercial modeling software UG is used to construct the small tongue angle, partitioned and large tongue angle volute based on the volute of the model pump. The ANSYS CFX is used to complete the numerical calculations. Comparative analysis of the hydraulic performance and pressure pulsation near the tongue of the three large volute pumps with different types of volutes are carried out. The results show that the efficiency of the partitioned volute plan decreases significantly under the over-load condition, and the pressure pulsation near the tongue is melted under the design condition. The small tongue angle plan is more suitable for large volute pumps.


Introduction
Large volute pumps are commonly used in large-scale water transfer projects, which are the key engineering equipment to solve the problem of long-distance and high-drop water transfer.The volute is one of the main components, which collects the medium and also plays the role of reducing the medium flow rate, so that the velocity energy is converted into pressure energy.Its main design methods include the constant velocity method and the method of conservation angular momentum of the flow.Chernobrova et al. [1] compared the volutes designed by the above methods respectively and found that the volute designed by the former method was more efficient.In order to get higher performance, many scholars have used various methods to optimize the design of volutes.Guo et al. [2] proposed a method for optimal design of the volute based on Radial Basis Function (RBF) neural network and Genetic Algorithm (GA) to optimize the efficiency and total sound pressure level of centrifugal pumps.Chen et al. [3] optimized the cross-sectional area of the volute by using intelligent algorithms to achieve the goal of high efficiency for double suction pumps.
The tongue is an important position of the connection between the spiral section and the diffusion section of the volute, which has a greater impact on the performance of the volute.Zhang et al. [4] conducted an experimental study on the effect of three different tongue cuts on the pressure pulsation of a centrifugal pump with a low specific speed, and found that the cut tongue has a significant effect on the pressure pulsation.Yang et al. [5] investigated the effect of radial gap between the impeller and volute tongue in a pump as turbine on the performance and pressure pulsations.It was found that the existence of a suitable clearance optimizes the pump performance and the intensity of pressure pulsation decreases as the gap increases.Aka et al. [6] investigated the effect of the tongue angle of the volute on the hemodynamic characteristics of a blood pump.Zhou et al. [7], on the other hand, investigated the effect of the geometry of the volute on the radial force characteristics of a centrifugal pump during startup.The partitioned volute is a special type of volute in which the tongue is integrated with the vane in the corresponding position.It is commonly used in pump turbines, and recently some enterprises have introduced it into the hydraulic model of large volute pumps, but there are few comparative studies on the specific application effects.In this study, the effects of three different kinds of volute on large volute pumps, including partitioned volute, are compared using a method of combining experiment and simulation.This paper provides a reference for the application of partitioned volute in large volute pumps.

Hydraulic model
In this paper, a large volute pump with a specific speed of ns = 235 is taken as the object of study, with a design flow rate of Qd = 920 m 3 /h, a head of H = 21 m, and a rotational speed of n = 1250 r/min.Figure 1 shows a three-dimensional model of the computational domain, which consists of inlet pipe, impeller, diffuser, volute, and outlet pipe, and the inlet and outlet pipes are appropriately lengthened to ensure the stability of the computation.Where the specific speed was calculated from equation (1).The positional parameters of the tongue were modified to change the volute type and the specific differences are shown in Figure 2. Plan A is the volute used on the model pump with a small tongue angle of 28.5°.Plan C is a model with a large tongue angle, which increases to 34°.Plan B takes the form of a partitioned volute, where the trailing edge of the diffuser vane near the tongue extends naturally forward until it is integrated with the tongue.

Mesh independence analysis
Different meshing strategies are adopted for different characteristics of different components.For the impeller and diffuser, TurboGrid is used for hexahedral structural meshing; for the volute, ANSYS ICEM is used for hybrid meshing with tetrahedral core; for the inlet and outlet pipe, ANSYS ICEM is used for hexahedral structural meshing.All wall surfaces are set up with boundary layer, the height of the first layer is 0.05mm, and the key wall surfaces such as the tongue are refined.The mesh of the main components is shown in Figure 3.
A total of five sets of grid nodes of 1.034 million, 4.072 million, 6.996 million, 10.155 million and 13.195 million are generated for the grid sensitivity analysis, and the obtained curves are shown in Figure 4 with the head as the monitoring index.When the number of grid nodes reaches 6.996 million, the change of head tends to level off, so this set of grids is selected for subsequent calculations.In this set of grid, the number of grid nodes of impeller, diffuser, volute, inlet and outlet pipe are 2.659 million, 2.555 million, 0.765 million, 0.755 million and 0.262 million, respectively.

Numerical calculation settings
The ANSYS CFX software was used to complete the steady and unsteady numerical calculations of the hydraulic model.The inlet boundary condition is the total pressure (1 atm), the outlet boundary condition is the mass flow (255.1 kg/s), and the convergence residual RMS is set to 10 -5 .The steady calculation iteration is 400 steps; the unsteady calculation time step is 0.0004s (the impeller rotates by 3°), and the impeller rotates by one cycle, which takes 120 steps and 0.048s, and the total calculation time is 0.72s (impeller rotation of 15 cycles), take the results of the last 5 rotational cycles of the calculations for analysis.Four monitoring points are set up near the tongue, named VC1, VC2, VC3 and VI1 in turn, and the specific locations are shown in Figure 5.The key hydraulic parameters are monitored during the calculation, and the calculation is considered to start converging when periodic changes occur.The SST k-ω two-equation eddy-viscosity model was used for the solution.

Experimental verification
The hydraulic performance test of the model pump was carried out on the closed-type test bench of Jiangsu Aerospace Hydraulic Equipment Co. to verify the accuracy of the numerical calculation results.The closed-type test system is shown in Figure 6.
The hydraulic performance data of the model pump under different operating conditions obtained from tests and calculations are compared, and the curves are shown in Figure 7.The efficiency, head and power curves are in good agreement.The relative errors of efficiency and head are all less than 5% under the design condition, at 1.42% and 3.5%, respectively.This indicates that the numerical calculations accurately reflect the internal flow characteristics of the model pump, and the results are reliable.

Comparison of hydraulic performance
The hydraulic performance curves of the model pump for different plans are shown in Figure 8.In the efficiency curve, the design condition (1.0 Qd) and part-load conditions (0.6-0.9 Qd) coincide with each other, and the difference is insignificant, the main difference is in the over-load condition (1.1 Qd), the efficiency of plan B decreases by 1.5% compared with that of plan A, and the efficiency of plan C also decreases by 0.37%, which is much smaller than that of plan B. And the efficiency of plan C is slightly higher than that of plan A in the design condition.In the head curve, the difference in most of the conditions (0.7-1.0 Qd) is also insignificant, only in the part-load condition (0.6 Qd) and over-load condition (1.1 Qd), there is a little difference.Under the 0.6 Qd condition, the head of plan C decreases, and under the 1.1 Qd condition, the head of plan B decreases.Overall, plan A still has the best hydraulic performance curve of the three plans.

Analysis of entropy production loss under the over-load condition
The distribution of entropy production rate (EPR) in the volute under over-load condition (1.1 Qd) of different plans is shown in Figure 9.In plan A, the large EPR region is mainly distributed near the tongue, toward the left side, and gradually decreases into the diffusion section; it is extended in the direction of flow in a long strip at the outlet of the upper and lower diffuser channels.In plan B, the area of large EPR region at the outlet of both upper and lower diffuser channels is enlarged, and there are new large EPR regions in the lower two channels near the tongue.The total energy loss increases significantly compared with that of plan A, which leads to a significant decrease in the efficiency of plan B under the over-load condition (1.1 Qd).In plan C, the large EPR region around the tongue is also significantly enlarged compared with plan A, which leads to a decrease in the efficiency of plan C under the over-load condition (1.1 Qd).

Comparison of pressure pulsations under the design condition
In order to clarify the variation of pressure pulsation, the pressure coefficient Cp is defined to characterize it and its expression is: Where ΔP is the difference between the time-varying pressure and the arithmetic mean of the pressure during the cycle, and u2 is the circumferential velocity of the impeller outlet.The peak-to-peak value ΔCp is the difference between the maximum and minimum values of the pressure coefficient Cp in one impeller rotation cycle, i.e. the fluctuation amplitude of the pressure coefficient.Calculated from the rated speed of the model pump, the shaft frequency fn = n/60 = 20.83Hz; the impeller blade passage frequency fBPF = Zim*fn = 125 Hz, which is 6 times the shaft frequency; and the diffuser vane passage frequency fVPF = Zdi*fn = 166.67Hz, which is 8 times the shaft frequency.
The pressure coefficient fluctuations during the last rotation cycle at each monitoring point in different plans are shown in Figure 10.Six full peaks and troughs can be observed at each point, and the waveforms of pressure coefficient fluctuations in plane A and C are more consistent, with plan C having smaller fluctuations.Plan B, on the other hand, has larger fluctuations at all points, with sharper peaks and valleys, and the fluctuation amplitude is substantially higher than that of plan A. In the frequency domain (Figure 11), the main frequency of each point under each plan is still dominated by the impeller blade frequency, and the other peaks are its octave, and the peak of the main frequency is the most in plan B, which indicates that the pressure pulsation status deteriorates in the vicinity of the monitoring point of concern for plan B under the design condition.The probable reason for this is that the vicinity of the tongue, due to its close to the diffuser outlet, is a low-pressure zone within the volute housing.Plan B separates this low-pressure area into two halves, so that fluid flowing through the same diffuser channel no longer passes through the connection and directly into the diffuser tube, but instead enters the beginning of the spiral channel.When the same amount of fluid passes through, the space available for passage is reduced, resulting in an increase in pressure, leading to a deterioration of the pressure pulsation on the right side of the tongue.This is also the reason for the deterioration of pressure pulsation on the left side of the tongue.

Conclusion
In this paper, the impact of three different types of tongue volutes on large volute pumps, namely, small tongue angle volute, partition volute and large tongue angle volute, is analyzed comparatively.The effect of the three types of volute on the hydraulic performance of large volute is mainly in the over-load condition (1.1 Qd).Under this condition, the highest efficiency is achieved by the small tongue angle volute plan, and the lowest efficiency is achieved by the partitioned volute plan, with a difference of 1.5%, while the hydraulic performance curves of the three pumps are almost the same under the rest of the operating conditions.The main frequency of pressure pulsation at the monitoring point of the three plans under the design condition is the impeller blade frequency, but the pressure pulsation status of the partitioned volute worsens, and the fluctuation of the pressure coefficient and the amplitude of the main frequency increase sharply.Therefore, from the point of pressure pulsation, the partitioned volute lacks obvious advantages in the application of large volute pumps, and the small tongue angle volute is more suitable.

Figure 1 .
Figure 1.3D model of computational domain.

Figure 2 .
Figure 2. Diagram of different types of volute tongue.The positional parameters of the tongue were modified to change the volute type and the specific differences are shown in Figure 2. Plan A is the volute used on the model pump with a small tongue angle of 28.5°.Plan C is a model with a large tongue angle, which increases to 34°.Plan B takes the form of a partitioned volute, where the trailing edge of the diffuser vane near the tongue extends naturally forward until it is integrated with the tongue.

Figure 3 .
Figure 3.The grid and details of the main components in the computational domain.

Figure 5 .
Figure 5. Distribution of pressure pulsation monitoring points.

Figure 7 .
Figure 7.Comparison of test and calculated hydraulic performance curves.

Figure 8 .
Figure 8.Comparison of hydraulic performance curves of different schemes.

Figure 9 .
Figure 9. Distribution of internal EPR in different plans under the over-load condition.

Figure 10 .Figure 11 .
Figure 10.Fluctuation of pressure coefficient in the last rotation period at the tongue in different plans.