Experimental Study and Analysis on Wear Characteristics of Mining Pumps Transporting Solid-Liquid Two-Phase Flows

Abstract

The use of mining pumps to transport solid-liquid two-phase flows results in solid particles causing varying degrees of wear in the flow-passing components of the pump. In this study, a computational fluid dynamics analysis was performed employing SST k-ω turbulence and discrete phase models to predict the wear characteristics of small two-stage mining pumps. The wear rates of the flow-passing components (first- and second-stage impellers and guide vanes) of the pump were simulated considering three solid-particle sizes under three flow conditions (low flow, rated flow, and high flow), with the highest wear rate occurring on the surfaces of the first-stage guide vane in the low-flow condition. Subsequently, a pump wear experiment was performed using an experimental pump with parameters and structure similar to the numerical model at the pump’s rated flow. The experimental results were compared to numerical results obtained using the same numerical method and wear model. The wear characteristics shown in the numerical and experimental results were consistent with each other. Thus, the numerical method used in this study can accurately predict the wear of flow-passing components in a mining pump, and the method was found to be suitable for the prediction of wear characteristics in mining pumps.

1. Introduction

During the transportation of solid-liquid two-phase mixtures, the flow channels of the pump are subjected to varying levels of wear. In addition, each particle (e.g., particle size) and operational (e.g., flow rate) parameter exerts a different effect on the wear characteristics of the flow-passing components of a pump. Therefore, studies regarding the internal wear characteristics of a pump transporting solid-liquid two-phase flows are necessary to improve pump design, and usage and achieve optimization [1,2,3,4,5].

Several factors can cause pump wear; however, the five primary factors that cause significant wear are slurry concentration, flow rate, angle of impact, particle size, and particle shape [6]. Tarodiya et al. reviewed the literature on pump wear characteristics from various perspectives, including the properties of the solid particles, flow characteristics of the slurry, and the concentration of solids, and consequently provided a useful reference for the optimization of pump design [7]. Dong et al. performed maximum stress computations to study the wear caused by coarse particles in the flow-passing components of a centrifugal pump and showed that the wear on the flow-passing components of a pump depends on particle size, particle shape, angle of impact, and flow rate [8]. Noon et al. performed numerical simulations to study the effects of slurry-particle impact velocity, mass concentration, and particle size on the wear of the flow-passing parts and casing of a centrifugal pump. They found that particle concentrations greater than 20% resulted in the wearing of the casing wall to gradually increase over time [9]. Shen et al. studied the effects of sediment particle size, particle shape, and particle concentration on the erosion wear of the impeller of a double-suction pump in the Jingtai Yellow River Irrigation Project and found that erosive wear primarily occurred around the leading edge (LE) and trailing edge (TE) of the impeller [10]. Huang et al. analyzed the impeller wear caused by solid particles, considering a variety of sediments with varying LE angles, and found that it mainly occurred on its working surfaces and rear cover [11]. Further, to predict the wear caused by solid-liquid two-phase flows on slurry pumps, Pagalthivarthi et al. studied the velocities and concentrations of solid particles in a radial cross-section of a slurry pump and performed simulations using a 3D slurry pump model [12,13]. Based on an analysis of impeller wear in centrifugal slurry pumps, Peng et al. proposed an optimal hydraulic design to minimize impeller wear [14].

In addition to numerical studies regarding the effects of particle size parameters and operating conditions on pump wear characteristics, many researchers have also employed experimental methods to study pump wear characteristics. However, the vast majority of these studies are qualitative in nature, and they aim to address pump wear problems by observing the state of wear zones inside the flow channels of a pump. It has been experimentally demonstrated that particle erosion is the primary cause of impeller damage in centrifugal pumps [15,16]. In a computational study by Peng et al. focused on the wear of heavy-duty slurry pumps, small flow rates were found to result in very unstable pump flow as well as significant backflow. In addition, they also found that an increase in particle concentration increased flow resistance, backflow, and local wall abrasion. These computational results were subsequently confirmed via field tests [17]. Walker et al. performed a comparison between the wear of slurry pumps in field and laboratory conditions, based on which specific wear conditions of a pump (local velocity, concentration, particle size, size distribution, and particle shape) were concluded to be important for further study to better understand pump wear problems [18]. Janssen et al. performed experiments using laser-manufactured microstructures to obtain optimal texturing parameters for positive displacement pumps and found that pumps designed with these parameters exhibited significantly improved wear performance [19]. Serrano et al. studied the effects of sediment concentration in Brazil’s Acre River on fluvial water pump impellers and performed ball-catering microscale abrasive wear tests to determine the erosive power of the sediments. Their results indicate that impeller wear increased almost linearly with an increase in sediment concentration [20]. Batalović studied the wear characteristics of slurry pumps under a variety of working conditions using a mathematical model of erosive wear in slurry pump impellers and performed tests to experimentally validate their approach [21]. Sugiyama et al. presented a numerical method for predicting the wear depth distribution on the blade of an impeller and validated the proposed method using experimental data [22].

The numerical methods and erosive wear models used in studies focused on pump wear can vary depending on the study parameters, and the models also differ in their range of applicability. Zhao used the discrete phase model (DPM) to analyze the wear caused by different flow rates and solid particle diameters in flow-passing parts and consequently obtained the wear rates of all flow-passing parts [23]. Wu used the dense discrete phase model (DDPM) to analyze the transient wear characteristics of centrifugal pumps, focusing on how each particle-related characteristic affected the wear of flow-passing parts in a centrifugal pump [24]. Eulerian multiphase flow models are frequently used to analyze pump wear characteristics associated with the transportation of solid-liquid two-phase flows, and their use has produced many useful insights [25,26,27,28]. Tarodiya et al. used an Euler–Euler model to analyze the effect of changes in the particle size distribution (PSD) on particle motion characteristics and the performance of centrifugal slurry pumps. Their results indicate that changes in the PSD affect the motion of particles in the pump flow channel, as an increase in the proportion of fine particles results in a decrease in the intensity of granular pressure, maximum granular viscosity, and head loss [29]. Peng et al. used the particle Eulerian–Eulerian multiphase model to study impeller wear in slurry pumps [30]. Further, Tarodiya et al. employed the Mixture and Eulerian–Eulerian multiphase flow models to study the wear performance of centrifugal slurry pumps during the transportation of solid-liquid two-phase flows and found that the experimentally observed effects of solids on pump performance were better predicted using the Eulerian–Eulerian model than the Mixture model [31]. In certain studies, particle and heterogeneous models were used to analyze the flows of slurry-pump rear impellers of various shapes and thus predict the wear characteristics of slurry pumps [32,33]. Cai et al. employed particle and heterogeneous models to study the effects of rear impeller shape on the wear characteristics of slurry pumps and thus predict the wear of slurry pumps [34]. Liao et al. used a two-phase particle flow model to analyze the effects of sand-particle size on the wear of impeller surfaces [35]. Yang et al. used the Particle two-phase and Finnie models to conduct a comparative analysis of the wear characteristics of pulp pumps; the former exhibited severe erosion of the front cover and impeller blade when the pump was operated at the rated flow, whereas the pump casing was severely eroded at low flow rates in the latter [36]. Liu et al. used the Finnie wear model to analyze the wear characteristics of the flow-passing components of deep-sea lifting pumps, which yielded the wear rates of each flow-passing component [37].

Thus, based on a review of the current literature regarding pump wear prediction, this study aimed to analyze pump wear characteristics associated with the transportation of solid particles, considering three different particle sizes using a mining pump under three different operating conditions (low, rated, and high flow). The particle trajectories and solids concentrations in each of the aforementioned conditions were obtained along with the wear rates of all flow-passing components (first-stage impeller, second-stage impeller, first-stage guide vane, and second-stage guide vane). In addition, qualitative wear experiments were performed using an experimental pump, and the observed wear inside the experimental pump’s flow channel was compared to that obtained using the numerical results to validate the applicability of the numerical method and wear model. The findings of this study are expected to be significant for the prediction of wear characteristics in mining pumps.

2. Methods: Three-Dimensional Wear Model and Numerical Method

2.1. Three-Dimensional Modelling

The main design parameter values of two-stage mining pumps adopted in this study were a rated flow rate Q_d = 20 m³/h, a single-stage head H_d = 15 m, a rated efficiency η_d = 64%, and a rated rotational speed n = 2860 r/min. The three-dimensional model of the two-stage mining pump is shown in Figure 1.

Since this article mainly analyzes the wear of the main flow-through components in the mining pump under the condition of solid-liquid two-phase flow, it mainly introduces the parameters and structure of the impeller and the guide vane. The geometric parameters of the impeller and guide vane of the mining pump are shown in

Table 1. Geometric parameters of the mining pump impeller and guide vane.

Impeller		Guide Vane
Parameter	Value	Parameter	Value
Inlet diameter/mm	60	Number of blades/piece	8
Outer diameter/mm	116	Wrap angle/°	90
Number of blades/piece	6	Inlet inner diameter/mm	120
Blade wrap angle/°	130	Inlet outer diameter/mm	145
Exit placement angle/°	28	Outlet inner diameter/mm	25
Outlet width/mm	12	Outlet outer diameter/mm	60

.

The hydraulic model of the mining pump impeller is shown in Figure 2a, and the hydraulic model of the guide vane is shown in Figure 2b.

2.2. Numerical Calculation Method

2.2.1. Meshing

In the numerical calculation, the calculation domain of the full flow field of the mining pump includes the inlet section, the first stage impeller, the first stage guide vane, the secondary impeller, the secondary guide vane, and the outlet section, as shown in Figure 3.

The entire calculation area is divided by structured meshes. Before calculation, ICEM software is used to generate 6 sets of meshes of different numbers, and the independence check is performed. The relationship between the number of meshes and the head of the mining pump is shown in

Table 2. Mesh independence verification.

Mesh Scheme Number	Impeller (Ten Thousand)	Guide Vane (Ten Thousand)	Inlet/Outlet Section (Ten Thousand)	Total (Ten Thousand)	Head (m)
1	9.1	24.9	13.2	94	31.71
2	15.2	42.0	22.8	160	30.84
3	20.0	55.1	29.9	210	31.22
4	29.1	80.2	42.9	300	30.35
5	51.5	142	76.0	530	30.36
6	66.0	184	101	700	30.31

.

It can be seen from Table 2 that when the number of meshes reaches 3 million, the head is 30.35 m; 5.3 million meshes correspond to a head of 30.36 m; subsequently, 7 million meshes correspond to a head of 30.31 m. It can be seen that when the number of meshes is greater than 3 million, the rate of change of the mining pump head is controlled within 1%, and the rate of change tends to be stable. Therefore, the number of meshes in the computational domain is finally determined to be about 3 million.

2.2.2. Numerical Calculation Method

Ansys Fluent 18.0 was used to perform steady numerical simulations of the computational domain of the mining pump. In the computational process, it is assumed that there is no energy or mass exchange between the particle phase and the fluid phase and that gravity is taken into account. The computation of the fluid domain adopts the SST k-ω turbulence model, the motion of the particle phase in the computational domain adopts the DPM model based on Euler–Lagrange coordinates, the discretization of the Navier–Stokes equation adopts the second-order upwind style, and the computational convergence precision is set as 10⁻⁴ in the numerical simulation.

Using the x-axis direction as an example, the equation for particle motion in the computational domain can be expressed as follows:

$$\frac{d{u}_p}{dt}=F_D\left(u-u_p\right)+\frac{g_x({\rho }_p-\rho )}{{\rho }_p}+F_V+F_P+F_x$$

(1)

where $F_D(u-u_p)$ is the drag force on a unit mass of solid-phase particles, as follows:

$$F_D=\frac{18\mu }{{\rho }_p{d}_p^2}\frac{C_D{R}_e}{24}$$

(2)

where, u is the fluid velocity (m/s), $u_p$ is the particle velocity (m/s), $\mu$ is the hydrodynamic viscosity (N·s/m²), $\rho$ is the fluid density (kg/m³), ${\rho }_p$ is the particle density (kg/m³), $d_p$ is the particle diameter (mm), and $g_x$ is the acceleration of gravity in the x-axis direction (m/s²).

2.2.3. Hydraulic Performance Test

To verify the correctness of the numerical model, we conducted hydraulic performance tests on the mining pump. In numerical calculations, different inlet velocities can be given to reflect the inlet boundary conditions of mining pumps under different flow conditions in order to obtain performance output curves under different flow conditions, as shown in Figure 4.

By comparing the performance curve obtained from the experiment, it can be found that the numerical simulation results are consistent with the experimental results, indicating the effectiveness of the numerical calculation method.

2.3. Boundary Conditions

The mining pump inlet adopts the form of a velocity inlet. The inlet velocity is calculated based on the inlet flow and the inlet through-passage cross-sectional area. At the same time, the solid-phase particles at the inlet face enter the computational domain axially, their entry velocity being the same as the fluid velocity—that is, there is no relative velocity. The outlet of the mining pump adopts the boundary condition in the form of outflow and defines the flow velocity weighting to be 1. All through-passage surfaces in the computational domain adopt non-slip boundary conditions, with the wall roughness defined as 0.5 based on the real blades of the model pump. The data transmission between the inlet section and the impeller and between the impeller and the spatial guide vane adopts the interface method. In the setting of the boundary conditions for solid-phase particles, the escape boundary condition is used for the pump inlet and outlet, and the reflecting boundary condition is used for the surfaces of other through-passage components.

3. Results: Prediction of Wear Characteristics in a Mining Pump

3.1. Wear Model

According to the existing literature, the wear empirical models proposed by the researchers are summarized, and as they have similar laws, they can all be summarized into the following equations [38].

$$R_{erosion}={\displaystyle \sum _{n=1}^{N_{particles}}\frac{m_{p}C(d_p)f(\theta )v^{b(v)}}{A_{face}}}$$

(3)

where:

$R_{erosion}$: erosion rate;

$C\left(d_p\right)$: particle size function of the solid-phase particles;

$\theta$: impact angle between the particle and the wall;

$f\left(\theta \right)$: function of the impact angle;

$v$: relative velocity between the particle and the wall, m/s;

$b\left(v\right)$: function of the relative velocity between the particle and the wall;

$A_{face}$: unit surface area of the wall, mm²;

$N$: the number of particles impacting the cell surface area;

$m_p$: mass of a single particle;

$d_p$: particle size;

$n$: number of particles.

After combining the above formula and the particle motion trajectory for analysis, the wear on the surface of the pump’s various through-passage components caused by the movement of solid-phase particles in the computational domain can be quantitatively predicted. The impact angle function, f(θ), can be expressed using a piecewise polynomial as follows:

When θ ≤ 15°,

$$f(\theta )=b{\theta }^2+c\theta$$

(4)

When θ > 15°,

$$f(\theta )=x{\cos }^2\theta \sin (w\theta )+y{\sin }^2(\theta )+z$$

(5)

The values of the constants in the above formula are shown in

Table 3. Values of parameters in the impact angle equation.

b	c	x	y	w	z
−13.3	7.85	1.09	0.125	1	1

b, c, x, y, w, and z are empirical constants in the impact angle function equation.

.

3.2. Computational Scheme

The numerical simulations were performed for varying particle sizes. According to Durand’s equation [39], computations involving solid-liquid two-phase flows should, under normal circumstances, be configured to ensure that the solid particles throughout the flow field have velocities greater than their settling rates and the critical velocity calculated using the Durand equation [39]. This prevents particle settling from affecting the computations. The computational scheme is shown in

Table 4. Computational scheme.

Flow Conditions	Particle Volume Concentration	Particle Density	Particle Size (mm)
0.65Qd	7.5%	1900 kg/m3	1, 3, 5
1.0Qd	7.5%	1900 kg/m3	1, 3, 5
1.3Qd	7.5%	1900 kg/m3	1, 3, 5

.

The study aimed to obtain the particle motion trajectories and concentration distributions of solid particles under low, rated, and high flow conditions, considering three different solid particle sizes. This is because these characteristics are closely related to the wear caused by solid particles in mining pumps. In addition, the effects of particle size on the wear rates of the flow-passing parts of a mining pump were analyzed as well.

3.3. Wear Predictions for the Low-Flow Condition

3.3.1. Analysis of Particle Motion Trajectories and Concentrations

The particle motion trajectories and solid concentrations simulated for a mining pump with three different solid-particle sizes in the low-flow condition are shown in Figure 5.

As evident, the solid particles moved at a uniform velocity in the axial direction into the impeller, and the work performed by the impeller on the particles caused their velocities to increase. However, the kinetic energies of the particles decreased slightly after they entered the guide vane. Moreover, this behavior was repeated when the particle entered the second-stage impeller and guide vane. Consequently, this trend increases the probability of collisions in this region and thus exacerbates wear.

According to the distribution of solid-particle concentrations in the blade-to-blade (B2B) section, the particles experience a transition in their velocity vectors between the impeller and guide vane, which results in high cross-sectional particle mass concentrations at the impeller-guide vane transition zone as well as near the outlet of the mining pump. This behavior was observed in the case of all particle sizes. Further, in the impeller region, the particles primarily contact the working surfaces of the impeller blades. However, with an increase in particle size, the high concentration zones moved away from the surfaces of the impeller. Furthermore, the area of the high mass concentration zone expanded with increasing particle size, particularly in the transition zone between the first-stage guide vane and the second-stage impeller. This implies that the throughput of a mining pump decreases when transporting large particles because they may clog the pump and thus decrease pump performance.

3.3.2. Wear Rates of the Flow-Passing Components

To quantify the wear of the flow-passing components, the average wear rates of the first- and second-stage impellers and guide vanes were calculated using area-mass weighted averages. Subsequently, the effects of particle size on these wear rates were analyzed, as shown in Figure 6.

In Figure 6, it is shown that wear intensity decreased with increasing particle size at the first-stage guide vane, second-stage impeller, and second-stage guide vane (i.e., all flow-passing parts except the first-stage impeller). Although the wear intensity trends were quantitatively different, in general, the wear was negatively correlated with particle size at the aforementioned flow-passing components. Moreover, the greatest wear rate was observed at the first-stage guide vane (13 × 10⁻⁸ kg/s∙m²) for a particle size of 1 mm, whereas the lowest was observed at the second-stage guide vane (4.8 × 10⁻⁸ kg/s∙m²) for a particle size of 5 mm.

The wear rate of the first-stage impeller increased with increasing particle size because the force of the impact between the LE and solid particles increased when particle mass (size) was increased with flow velocity maintained at a constant value. At other flow-passing components, an increase in particle size resulted in a reduction in the work done by the impeller on the particles for the same energy, which then reduced the increase in radial velocity. Furthermore, for an unchanged volume fraction, an increase in particle size reduces the number of particles. Consequently, both of these effects reduce wear.

3.4. Wear Predictions for the Rated-Flow Condition

3.4.1. Analysis of Particle Motion Trajectories and Concentrations

Figure 7 shows the particle motion trajectories and concentrations that were simulated for a mining pump transporting solid-liquid flows at its rated flow rate, for three different particle sizes.

As evident, the increased inertia at the rated flow (compared to the low-flow condition) led to significant particle deposition on the working surfaces of the impeller blades and guide vane. Further, the larger the particle size, the more obvious the deposition. In addition, the distribution of solids in the flow channel of the guide vane also became more uniform with an increase in particle size.

3.4.2. Wear Rates of the Flow-Passing Components

Figure 8 illustrates the average wear rates of the first-stage impeller, first-stage guide vane, second-stage impeller, and second-stage guide vane in the rated-flow condition for three different particle sizes.

As shown, the dependence of wear intensity on particle size in the rated-flow condition is similar to that in the low-flow condition. With an increase in particle size, the wear of the first-stage impeller increased; however, the wear at the three other flow-passing components decreased. Therefore, wear was positively correlated with particle size at the first-stage impeller but negatively correlated with other flow-passing components. Further, in the rated-flow condition, the greatest and lowest wear rates occurred at the first-stage guide vane when the particle size was 1 mm (11.7 × 10⁻⁸ kg/s∙m²) and the second-stage guide vane when the particle size was 5 mm (5.8 × 10⁻⁸ kg/s∙m²), respectively.

3.5. Wear Predictions for the High-Flow Condition

3.5.1. Analysis of Particle Motion Trajectories and Concentrations

Figure 9 illustrates the simulated particle motion trajectories and solids concentrations of a mining pump operating in a high-flow condition for three different particle sizes.

As shown, the distribution of particles in the guide vane channels became more uniform with increasing particle size, and the intensity of wear on the guide vanes’ blades decreased with increasing particle size. In addition, the concentration of solids in the stator-rotor region of interference between the outlet of the first-stage impeller and the inlet of the first-stage guide vane increased with increasing particle size.

3.5.2. Wear Rates of the Flow-Passing Components

Figure 10 shows the effects of solid-particle size on the average wear rates of the first- and second-stage impellers and guide vanes in the high-flow condition.

The wear intensity responses of the flow-passing components to increasing particle size in the high-flow condition are similar to those in the low- and rated-flow conditions. In addition, in the high-flow condition, the greatest wear occurred in the first-stage guide vane for a particle size of 1 mm (11.6 × 10⁻⁸ kg/s∙m²), and the lowest wear was observed in the first-stage guide vane for a particle size of 5 mm (6 × 10⁻⁸ kg/s∙m²).

4. Experiment: Wear Experiment Verification

To validate the applicability and accuracy of the wear model used in the numerical calculations (see Section 3.1), a pump wear experiment was performed using a mining pump. Subsequently, the experimental results were compared with the results obtained from a numerical simulation performed with identical boundary conditions.

4.1. Experimental Methodology

A two-stage mining pump with the parameters and structure described in Section 2.1 was constructed. Figure 11 shows a photograph of the experimental pump.

The experiment system of this experiment pump is composed of an experiment pump, a water tank, a regulating valve, a pipeline, a data measurement system, etc., as shown in Figure 12.

This experiment was conducted to observe the occurrence of real in the erosive wear zones of the experimental pump. To this end, regularly shaped solid particles with a particle size of 3 mm and a concentration of 7.5% were used, and the experiment was performed at the pump’s rated flow (20 m³/h). However, prior to the experiment, the impeller and guide vane flow channels were painted to the same thickness using water-based paint. Consequently, the experiment was performed after the paint had completely dried, and the wear at each flow-passing component was ascertained by observing the effects of the solid particles on the paint at the wear zones. Thereafter, the experimental results were compared with those of the numerical simulation that was performed with the same boundary conditions. Photographs of the impeller and guide vane that were acquired before and after they were painted are shown in Figure 13 and Figure 14.

The procedure of the experiment was as follows: (1) The experimental pump was installed in the experimental system, and a no-load test was performed to ensure that the electric motor and data acquisition system were functioning normally. The wear experiment was conducted after a normal operation was confirmed. (2) The impellers and guide vanes were disassembled, and the inlet, impellers, and guide vanes of the pump were wiped clean. Thereafter, a water-based paint was applied up to the same thickness. After the paint dried off, the experimental pump was reassembled. (3) Clearwater and the experimental sand were added to the water tank, with the latter being added up to a concentration of 7.5%. The experimental pump was then installed in the experimental system, and the connections of the pipes and valves were checked for correctness. (4) The power switch on the master console was turned on, and the “pump wear experiment” program was executed. The regulating valve was used to adjust the inlet flow to 20 m³/h, and the experimental data was recorded for 2 h. Subsequently, the power switch was turned off to terminate the experiment. Following the experiment, the impellers and guide vanes of the experimental pump were disassembled to observe the state of the water-based paint at the wear zones of the inlet, first-stage impeller, first-stage guide vane, second-stage impeller, and second-stage guide vane. Finally, photographs were acquired before the experimental site was tidied up. Photographs of the experimental site are shown in Figure 15.

4.2. Comparison between Experimental and Numerical Results

The wear characteristics of the experimental pump were numerically simulated using the numerical method and wear model described in Section 2, with boundary conditions identical to those of the pump wear experiment. The wear at the inlet and first- and second-stage impellers and guide vanes of the experimental pump were simulated, and the results obtained were compared to the experimental results, as shown in Figure 16.

In Figure 16a, it is shown that the numerically simulated wear is generally consistent with the experimentally observed wear, with some slight differences. At the inlet of the experimental pump, the solid particles caused severe wear at the radial inlet. This is because the radial inlet is the first part of the pump that comes into contact with the particles. Thus, owing to the particles being still fully intact at this stage and the particle collisions occurring at very high angles, severe wear occurred at this part of the pump.

From the comparison results of the first-stage impeller in Figure 16b, surface wear mainly occurred near their LE and TE. At the LE, solid particles moved in the axial direction into the impeller LE, and the rotations of the impeller resulted in their corresponding velocity changing from axial to radial. Consequently, this resulted in high-impact angles and collisions with the LE, thus causing severe wear at this location. The wear at the TE was primarily because the work done by the impeller increased the kinetic energy of the particles, whose tangential velocity was maximal at the TE, thus exacerbating wear at the TE. Moreover, the type of wear that occurred here was mainly abrasive wear.

From Figure 16c, it can be seen that the wear on the second-stage impeller’s surfaces was worse than that on the corresponding surfaces of the first-stage impeller because the particles experienced two rounds of acceleration by the second stage, and the resulting increase in particle kinetic energy also increased impact damage on the surfaces of the second-stage impeller’s blades.

However, on closely examining Figure 16b,c, minute differences between the numerical and experimental results in the wear zones of the impeller blades can be observed. In the numerical simulation, the boundary of the pump inlet was configured such that the particles entered the pump with a uniform distribution. Thereafter, once the particles contacted the LE of the impeller blades, their velocities changed from axial to radial. However, the length of the impeller blades was insufficient for the axial velocities of the particles to completely change into radial velocities, and thus, the simulated wear at the LE of the impeller blades was small and indistinct. Yet, after the particles were accelerated by the impeller, a large number of impacts occurred between the particles and the TE of the impeller blades, which greatly increased wear at the TE. In contrast, in the experiment, the particles did not enter the pump inlet with a uniform distribution; therefore, after the motor was turned on, the particles driven by the rotations of the impeller caused significant wear at the LE of the impeller blades.

In Figure 16d,e, it may be observed that the most worn areas of the guide vane blades were located near their LE and that the first-stage guide vane exhibited worse wear than the second-stage guide vane at all corresponding locations. This is because the rated flow is the optimal operating point of the pump. After the solid-liquid two-phase flow has been guided and worked on by the first-stage impeller, first-stage guide vane, and second-stage impeller, the particle flow entering the second-stage guide vane becomes stable, which reduces the number of collisions between the particle and the second-stage guide vane’s surfaces and thus reduces their wear.

5. Conclusions

In this study, a computational fluid dynamics (CFD) study was performed employing the SST k-ω turbulence and discrete phase (DPM) models to predict the wear characteristics of small two-stage mining pumps. It also outlined potential methods for developing a predictive model for long-term wear prediction taking into account factors such as material degradation, cumulative damage accumulation, and maintenance schedules. This would allow better planning and management of pump maintenance activities to minimize downtime and maximize equipment life. In addition, experimental verification studies have also been carried out. The following conclusions were drawn:

(1)

Numerical simulations were performed to compare the wear on the impeller and guide-vane surfaces under several conditions. During the transportation of small particles, the first-stage guide vane always exhibited the greatest surface wear, at all flow rates. However, during the transportation of large particles, the second-stage guide vane exhibited the least surface wear in the low-flow and rated-flow conditions, whereas in the high-flow condition, the first-stage guide vane exhibited the least surface wear. Therefore, the surface strength of the first-stage guide vane should be prioritized in the design and construction of small, two-stage mining pumps.

(2)

In the pump wear experiment, which was performed at the pump’s rated flow, severe wear was observed at the radial inlet of the pump, as well as the LE and TE of the first-stage impeller and second-stage impeller. Moreover, the surface wear on the second-stage impeller was more severe than that on the first-stage impeller. Hence, the flow-passing components must be strengthened at these locations.

(3)

A comparison was performed between numerically simulated and experimentally observed wear characteristics, which showed that the numerical and experimental results are consistent with each other. Therefore, the numerical method used in this work can accurately predict the surface wear of flow-passing components in mining pumps.

(4)

As mining pumps work in complex operating environments and comprise several flow-passing components, their particle motions are highly complex. When a mining pump is used to transport solid-liquid two-phase flows, the solid particles cause varying degrees of wear in the flow-passing components. However, by predicting the wear characteristics of mining pumps using an appropriate numerical method and erosive wear model, pump designers and manufacturers may use the insights gained from these predictions to selectively strengthen the parts of the flow-passing components that are expected to experience significant wear. Thus, the methods proposed in this study are expected to function as a suitable numerical method and wear model that is well-suited for this purpose.