A numerical study of clogging analysis in submersible drainage pump

Erosion of material surface due to collisions with solid particles has become a challenge to several engineering fields. Therefore, it is necessary to effectively design radial flow pumps that mitigate the sediment particles’ clogging effect from the system performance. This study aims to analyse and identify the clogging development of a newly designed submersible pump using computational fluid dynamics (CFD) techniques. The particle trajectory was accounted Tabakoff–Grant erosion model for predicting the clogging impact. The experimental pump performance results obtained by this study were used to validate the numerical results of the submersible pump. Pump results revealed a good agreement between the two marks, especially the head of the pump was obtained at a flow rate of 0.165 m3/min. When considering the motor power factor of 0.78, the pump efficiency was found to be 46%, and the test result was well-matched. Also, for combined cavitation-erosion, the performance characteristics of erosion rate density increase linearly. At a flow rate of 0.16 m3/min, the erosion rate density of the original model was more significant than the revised model. Furthermore, compared to the original model significantly reduced the adverse clogging effects on the impeller blades of the revised model.


Introduction
A submersible pump is a type of radial flow pump, the energy is added to the liquid when the fluid is flowing in radial directions [1].An essential problem during the radial flow pump operation is cavitation and cavitation-sand erosion [2,3].Cavitation-sand is difficult to avoid for many technical applications but must control at all operating conditions.Therefore, one needs detailed insight into the mechanisms governing cavitation-erosion phenomena.In addition, abrasives in the water may wear out the submersible pump.As such, erosion of material surface due to collisions with solid particles has become a challenge to several engineering fields.It is then necessary to analyze and identify the effect of the sediment particles to predict the erosion rate correctly.Therefore, additional study and refinement are required to investigate the relationship between particle movement and pump erosion.
In the literature, cavitation is addressed as a significant issue with bigger centrifugal pumps when the interval between the onset of the case and failure might be small [4].If a centrifugal pump operates below total capacity, this is energy inefficient.Many experiments have been carried out throughout the years to study the physical aspects of cavitation [5].Since then, much theoretical and experimental research has also been conducted.Both practical and computational studies have been done to avoid the adverse effects of cavitation.However, empirical studies are unsuitable for understanding cavities correctly and are not cost-effective.
On the other hand, computational fluid dynamics that predict the position and extent of the cavity are mostly adequate to be assimilated into the design of turbomachines.Hence, technology for the well-defined prediction and the estimation of cavitation and cavitation-erosion are meaningful in developing fluid devices such as centrifugal pumps to reduce these adverse effects.A transport equation model (TEM) has been revealed to consider the cavitation dynamics in an adaptable manner which have been applied in CFD software such as the Zwart model, the Kunz model (Kunz et al., 2000), the Singhal model (2002), etc. [6][7][8].
In addition, solid particles can also influence cavitation erosion.Generally, silt erosion is the leading cause of component damage.Combined silt erosion carried vapor bubbles into the largepressure region, producing cavitation-erosion [9].The research significantly reduces abrasion, increases service life, and promotes design for the multiphase flow pump.Only a few studies have been carried out on solid-liquid two-phase flow, but there has yet to be a study on cavitation-sand erosion in the pump.Shi and Wei [10] had conducted on solid-liquid turbulent flow analysis in a centrifugal pump numerically, and found the main region of the deterioration of the impeller and casing.Analysis of the energy exchange mechanism of a screw centrifugal pump investigated in sand particle-liquid conditions and acquired the power capability of the impeller in diversity phase [11].The generic algorithm and decision tree have been applied to find out an optimized anti-cavitation blade shape of impeller of the centrifugal pump [12].Zhu J. et al. [13] explored a mechanistic model for enriching the performance of gas-liquid flow in submersible pumps.Experimental studies on pump clogging of submersible sewage pumps have been conducted [14] to ascertain the coincidence between the execution of an anti-clogging and pump.By using both experimental and computational methods, the clogging process of fibrous substances in a pump has been investigated [15].Hence, mitigating cavitation and the clogging effects of the submersible drainage pump is still unresolved.
Therefore, in this study, numerical simulations of the submersible pump are performed on Euler-Euler two-phase model to determine the cavitation performance and erosion rate.The effects of cavitation characteristics and erosion in the pump are described in detail.

Geometrical model and meshing
An assembly drawing of a vertical semi-open type submersible pump (model-DWE-08B) was prepared to analyze its performance, as shown in Figure 1.The pump company provided the design parameter of the suction stand of the pump, and ANSYS ICEM-CFX (15.7) software generated the 3D geometry of the suction stand.The specifications of the submersible pump head were 10 m, the fluid stream rate was 0.16 m 3 /min, the rotational speed was 3450 rpm, and the number of impeller blades was 2, respectively.
The submersible pump has a complex internal geometry, so unconstructed prism-tetrahedral meshing grids were constructed.Figure 2 illustrates a submersible pump meshing grid.Different model mesh sizes were taken into consideration to improve the numerical accuracy [16].The pump pressure head was the criterion used to check the grid reliance of the pump, as shown in Figure 3.According to the grid reliance study, mesh 2 had a maximum difference of 0.56% in pressure head compared to mesh 4. Finally, mesh 2 was chosen to simulate the entire pump since its Y+ value fell inside the acceptable range for prism-tetrahedral grids [17,18].The total meshed grids were 272,563 nodes and 1,374,829 elements.

Governing equations and boundary conditions
The continuity and momentum equations served as the foundation for numerical analysis of the flow [18,19].The k-ω based shear stress transport model took the turbulent shear stress, which also made exact predictions of the initial and significant flow separation under unfavorable pressure gradients [19,20].In this work, the boundary conditions used for the simulation were mass flow rate as inlet and static pressure outlet.All boundary walls were anticipated to be smooth and no-slip.The frozen rotor was selected at a given rotational speed for steady-state, incompressible flow analysis in the rotating and stationary interface.The entire pump computational domain is shown in Figure 4.A highresolution-scheme was used to discretize the convective term of the governing equation, and SIMPLE algorithms were taken into account in solver control.The velocity and pressure residual value was 1x 10 -5 controlled by convergence criteria.
In cavitation, the inlet boundary's total pressure and mass flow rate are imposed at the outlet boundary.The Rayleigh-Plesset cavitation model was considered to predict and understand the cavitation phenomena.For numerical stability, the steady-state without cavitation solution was considered an initial condition for the steady-state with cavitation.The saturation vapor pressure was set to 3169 Pa at a water temperature of 25 °C.The wastewater has a specific gravity of roughly 2.65 and is made up of various sizes and particles, including clay, silt, sand, and gravel.Quartz was therefore chosen as the particle for investigating sand erosion and cavitation-sand erosion which has a 2650 kg/m 3 density, and its molar mass is 60.08 kg/kmol [17].The minimum diameter of the particle is 15 micron, maximum particle size is 2 mm, and mean diameter of the particle is 0.1 mm.For the duration of the pump's passage, the mass flow rate and pressure boundary conditions at the input and outflow (0 Pa) were taken into consideration.The Schiller-Naumann model considers for estimation of interphase forces comply with the particles.Tabakoff-Grant erosion model was adapted in ANSYS-CFX to estimate erosion phenomena.

Experimental results
The experiment and simulation data were compared to confirm the validity of the DWE-08B pump analysis results. Figure 5 shows the pump head, efficiency, and shaft power performances for different operating flow rates at a constant rotational speed of 3450 rpm.It is observed that when the specific point was 0.165 m 3 /min, the total head was 10 m, and the overall efficiency was 32.14% which was not the same as the pump efficiency.As demonstrated, for a specific flow rate, there was an adequate correlation between the two results, especially the pump head.This is because the test pump was in the water and couldn't measure the torque.Therefore, the power factor of the pump motor must be considered to compare it with the test results.Also, in this research, we did not regard the mechanical loss between the submersible pump's bearing and shaft [21].

Cavitation phenomena
This study looked into the cavitation phenomena for various casing and impeller shapes.For different pump types with a discharge rate of 0.16 m 3 /min, the estimated values of total head versus NPSH are shown in Figure 6.The head drop lines obtain by progressively lowering the suction pressure.The graph shows that the head drop propensity for the original and revised models (casing and impeller designs) was the same.It has also been noted that vapor bubbles gradually formed as the NPSH was decreased, causing reductions in the total head.A 3% head drop was taken into account for cavitation conditions based on these findings.Then, NPSH3% was taken into account for various pump models.NPSH3% for the original impeller, casing 2, impeller model 1, and impeller model 1 with casing 2 were 1.334 m, 1.282 m, 1.206 m, and 1.268 m at a stream rate of 0.16 m 3 /min.Table 1 compares the actual and impeller model 1's performance.Here, impeller shape change model 1 vanes were created diagonal (obliquely shape) form the actual model.From Table 1, the head of impeller model 1 reduces by approximately 0.23 m to that of the original model.Moreover, it is shown 0.13 m and 5.24% improvements in the NPSH3% and efficiency, respectively, compared to the original model.Therefore, the impeller model 1 design produced mostly stable flows in the impeller pathway.

Erosion effects on performance
The computational research was regarded as a solid-liquid phase flow in this study.Various sand particle influx rates ranging from 0.1 to 1 kg/s in the different pump models, were taken into consideration to check out the clogging effect on erosion.The distribution of particle inflow rates on the impeller blade is shown in Figure 7.Moreover, it has been found that the region of erosion rate density increases along with the particle inflow rate.On the impeller blades, the particle concentration distribution is uniform.As the inflow rate increased, the region where erosion rate density shows a more excellent value, but the erosion rate density is reduced for impeller model 1 with the comparison of the original impeller.The erosion rate was observed at first on the suction side of the blade's leading edge.This can be explained by the fact that the impeller blade's tip edge side has a more significant relative velocity.At this point, cavitation erosion had a significant impact on the flow channel and could be generated by pitting the impeller blade.

Clogging effects on performance
Water typically transports bubbles into higher-pressure regions in a cavitating circumstance, but in a cavitation-erosion scenario, water and sediment transport the vapor bubbles into the high-pressure zone [22].Consequently, this research explored under combined cavitation-sand erosion at four distinct particle influx rates ranging from 0.1 -1 kg/s at a discharge rate of 0.16 m3/min.The distributions of the impact of the sand volume fraction on the impeller blade are shown in Figure 8. Figure 8 illustrates that the region of the associated clogging effect from the leading-edge suction side grew as the erosion rates increased.In contrast to the original model, the flow pattern of the impeller model 1 has been changed from the front edge on the suction side to the rear edge, with minimal effect on the impeller blade's trailing edge.The leading-edge suction side of the pump was where erosion first began to spread, and it was discovered that the distribution of sand volume was not uniform.The flow channel was also seen to be almost clogged by erosion.For impeller model 1, the erosion rate density was comparatively lower than the original model and could reduce the clogging effect.

Conclusion
This study looked into enhancing a submersible drainage pump's performance and determining the clogging effect.When the submersible pump was subjected to cavitation at a high flow rate, a low head was experienced.Both the initial model and the shape-change models propagated on the impeller blade.It was discovered that impeller model 1's suction performance had greatly increased over the original model over the entire flow rate.Additionally, it demonstrated that NPSH3% had increased from the original model by around 0.13 m.Furthermore, the leading edge of the impeller was discovered to be the primary area of erosion.Additionally, the erosion rate density of impeller model 1 was found to be lower than that of the original model, according to the results.According to cavitation-erosion data, the flow distribution channel for impeller model 1 had moved from the suction leading edge to the pump impeller's trailing edge, and the erosion impact had not occurred at the blade trailing edge.As a result, it was judged that impeller model 1 had a reduced rate of erosion and could minimize the clogging impact.Further study should be considered the experimental study of the cavitation and erosion performances for the validation of the numerical results.

Figure 3 .
Figure 3.The pump head versus the number of meshing nodes.

Figure 4 .
Figure 4.The computational domain of the submersible pump.

Figure 5 .
Figure 5. Experiment and computational performances of the submersible pump (a) Head vs. flow rate (b) Efficiency vs. flow rate and (c) power vs. flow rate.

Figure 6 .
Figure 6.Head versus NPSH for various pump models at 0.16 m 3 /min.

Figure 7 .
Figure 7.The comparison of erosion rate density between original and impeller model 1.

Figure 8 .
Figure 8.The comparison of sand volume fractions on the impeller blades.

Table 1 .
Performance comparison according to impeller and casing shapes.