Experimental study on multi-condition of axial flow pump in the whole flow field based on piv test

To study the internal flow characteristics of the axial flow pump in the whole flow field under different flow conditions, particle image velocimetry technology was used to measure the steady and unsteady flow fields of the axial flow pump, analyzed the velocity field distribution at the impeller inlet, impeller inner, guide vane inner, and guide vane outlet of the pump under different flow conditions (Q/Q des=0.8, 1, 1.1). Through the test and data analysis: under the three flow conditions, the velocity distribution trend of the impeller inlet of the axial flow pump is consistent and basically along the axis direction; there is a large velocity gradient in the flow field inside the impeller, and there is a significant acceleration near the hub trend, and there is a high-speed concentrated area; the velocity distribution trend in the guide vane is basically the same, and the fluid velocity changes more uniformly at the rated flow rate, and the flow tends to be more stable; a local high-speed area appears at the outer edge of the guide vane outlet, but in the second half The flow velocity in the pipeline is relatively uniform. The test results can provide a reference for the internal flow characteristics and performance optimization design of the axial flow pump.


Introduction
Axial pumps are widely used in agricultural irrigation, urban flood control and drainage, cross-basin water transfer and high-performance ship jet propulsion due to their high flow rate and low head [1,2].With the continuous promotion of ships, South-North Water Diversion and other projects, axial flow pumps play an important role in the construction of the national economy.Inside the axial pump, the flow state of the fluid directly affects the work done by the impeller and the efficiency of pressure energy conversion.Therefore, in-depth investigation of the flow velocity distribution in the whole basin of the axial flow pump is of great significance to the design of axial flow pumps.
The current methods for studying the basin-wide flow field of axial pumps are mainly numerical calculations and experimental tests.Although numerical calculations have become one of the important means to improve the performance of axial pumps[3], the reliability of the results of optimised design in the face of multi-case, multi-objective and multi-constraint axial pumps is still worth discussing.Moreover, numerical calculations can only be used as a means of simulation[4-6], and the accuracy of the results still needs to be verified through tests.Existing means of flow field testing inside impeller machinery include PIV (Particle Image Velocimetry) and LDV (Laser Doppler Velocimeter) [7,8].The principle of LDV is using the Doppler effect to measure the velocity of a fluid.However, the measurement accuracy of LDV is easily disturbed by the flow state of the medium[9].And it is not possible to obtain all the statistical information of the flow field in a certain space in a single experiment.And PIV as a non-contact measurement method[10] is more advantageous in acquiring information on a wide range of flows.PIV is suitable for high speed and complex flow systems in the whole basin of axial pumps, which can obtain the instantaneous vector flow field with high time resolution and high test accuracy at the same time [11,12].
Therefore, in this paper, the PIV test technique is used to measure the axial cross-section flow fields at different flow rates at the impeller inlet, impeller inner, guide vane inner, and guide vane outlet of the pump, and then analysing the effect of flow variations on the internal flow law of axial pumps from the velocity vector field, so as to reveal the mechanism of unstable operation of axial pumps.After grasping the real flow field state, it could provide a reference for the performance optimisation design of axial flow pumps.

Physical model
The test takes the axial flow pump model with diameter D = 300mm as the research object.It is designed for a flow rate of Q des and a rated speed of n des = 1450rmin -1 .Specific parameters are shown in Table 1.

Specific speed n s =420
A physical drawing of the axial flow pump model for the PIV test is shown in Figure 1.In the test, the impeller inlet to the guide vane outlet tube section of the model was made of Plexiglas for the convenience of the camera.And its shape is square on the outside and round on the inside.The test section has an inlet line diameter of 300 mm and an outlet line diameter of 220 mm.

Experimental methodology
The particle image velocimetry system of the test mainly includes Vlite laser system, Speed Sense cross-frame camera, Dynamic Studio image acquisition and analysis software, synchronisation controller, and photoelectric encoder.The test was carried out using aluminium profiles to build stands.Taking the internal flow field of the impeller as an example, the position of the laser lens and camera is shown in Figure 3.The laser lens is fixed directly above the test section and the slice light source is directed vertically into the axial section of the axial pump.The laser lens can be manually adjusted left and right along the axis.The Speed Sense cross-frame camera is mounted on the profile outside the test section.And in order to eliminate the errors caused by optical refraction, special water prisms are used to meet the flow field test requirements.

PIV experiment
In this paper, the whole basin of the axial pump model is divided into four places for testing.The axial cross-section flow fields at the impeller inlet, impeller inner, guide vane inner and the outlet of the guide vane.This four places are obtained under the conditions of 0.8Q des , 1.0Q des and 1.1Q des , respectively.Where the impeller inlet, guide vane inner, and guide vane outlet are steady flow fields and the impeller interior is an unsteady flow field.To obtain stable time-averaged velocity field information, a motor encoder is used to synchronise the triggering of the laser to take phase-locked pictures of the test section when taking pictures of the internal flow field of the impeller.In the test, 2000 frames of images were taken for each test section and 500 of the better velocity fields were selected for time-averaging processing to get the time-averaged velocity field.The four test shaft sections and tracer particle effects are shown in Figure 4.

Flow field distribution of the impeller inlet shaft section
The flow field distribution of the impeller inlet shaft section is shown in Figure 5.The horizontal coordinate X is the axial distance of the shooting section and the vertical coordinate Y is the radial distance of the shooting section.It can be seen that the distribution of the flow field under different flow rates is basically the same, the incoming flow along the contour of the front guide cap to the vicinity of the inlet side of the impeller, the flow field velocity gradually increases, and a high-speed region is formed at the top of the inlet side of the impeller.This is due to the fact that at the same rotational speed, along the inlet side of the impeller radially outward, which its circumferential velocity is increasing, and the acceleration effect on the incoming flow is more pronounced.So the velocity of the flow field at the impeller hub is increasing towards the outer edge of the impeller inlet side, which satisfies the design requirements of the velocity moment distribution at the impeller inlet side.
At the same time, a low velocity zone exists at the top right of the graph, and this zone persists as the flow increases.According to the characteristics of axial flow pumps, it was known that the impeller inlet is prone to reflux in the case of small flow rate.However, the backflow phenomenon persisted in the three flow conditions.So it is analysed that a gap in the flange connection between the test section and the front end line along the direction of incoming flow, causing the flow separation, and resulting in a turbulent velocity distribution with the effect of the end-wall boundary layer.However, the thickness of the boundary layer decreases gradually and the acceleration process becomes more obvious with the increase of flow velocity.
Comparing the different flow rates, we can see that the maximum inlet velocity increases from 5.99ms -1 to 7.52 ms -1 .The low-speed region of the upper end wall and the impeller inlet hub continues to decrease while gradually shifting to the high-speed region.The velocity distribution at 1.1Q des is dominated by the high-speed zone and is more evenly distributed.The difference between the highest and lowest speeds under the three flow conditions is about 1 ms -1 .
(a) 0.8Q des (b) 1.0Q des (c) 1.1Q des Figure 5. PIV measurement results of impeller inlet shaft section.For the purpose of subsequent analysis.The absolute velocity magnitude of the flow field is expressed as normalised using the impeller inlet velocity.The velocity values within the rectangular region were extracted for averaging at the same locations in the mainstream region shown in Table 2.The inlet flow velocities of the impeller under 0.8Q des , 1.0Q des , and 1.1Q des were calculated as 5.37 ms - 1 , 6.40 ms -1 , 7.29 ms -1 respectively.
Table 2.The inlet flow velocities of the impeller.

Flow field distribution of the impeller inner shaft section
The flow field distribution of the impeller inner shaft section is shown in Figure 6.The velocity gradient inside the impeller varies more significantly.Velocity distribution increases in steps from the inlet to the outlet of the impeller, and deflects towards the outlet of the impeller hub, forming a region of high velocity.This is due to the fact that the laser sheet light source is a vertical cross-section through the axis of the impeller.So at the inlet of the test section, which is near the pressure surface, while at the outlet the test section is located near the suction surface.The flow velocity increases gradually from the pressure surface to the suction surface, which is consistent with the results analysed by the potential flow theory.
As the flow rate increases, the energy gained by the fluid inside the impeller continues to increase.The low-speed zone near the hub of the impeller inlet is decreasing, and the high-speed zone is gradually travelling upwards with axial acceleration.The flow field is more uniformly distributed along the axial direction at 1.0Q des , but the impeller inlet flow field distribution is more turbulent at 1.1Q des .

Flow field distribution of the guide vane inner shaft section
The flow field distribution of the guide vane inner shaft section is shown in Figure 7.As the guide vane is fixed and immobile, its flow has no relative flow, only absolute and steady flow.So the velocity distribution trend of the fluid is basically the same.Inside the guide vane, the fluid velocity increases gradually in the axial direction.This is due to the fact that in order to satisfy the optimal flow channel area change pattern, the flow path in the guide vane begins to shrink, resulting in reduced overcurrent area.So the circumferential induced velocity is eliminated while the fluid acceleration occurs.
Under different flow rates, compared to the unsmooth acceleration process at 0.8Q des and the sharp transition of the flow field at the outlet of the guide vane at 1.1Q des , the rectification effect of the guide vane on the static-dynamic turbulent flow field is more obvious at 1.0Q des .Meanwhile, the fluid velocity changes more uniformly and the flow tends to be more stable.

Flow field distribution of the guide vane outlet shaft section
The flow field distribution of guide vane outlet shaft section is shown in Figure 8.In the left half of the flow field, the flow velocity of the straight section of the pipe is more uniform, and the fluid flows smoothly outward in the axial direction.However, a localised region of high velocity exists in the vicinity of the shrinkage runner.The velocity distribution in the region tends to decrease sharply from the centre of the region to the periphery.This is due to the fact that the number of guide vane blades is seven, which causes the crowding coefficient larger.In addition, the guide vane flow path is shrinking, while the presence of the hub diameter results in a small effective overflow area, resulting in constant acceleration of the fluid at the outlet of the guide vane.
However, after the fluid transitions from the shrinkage section to the straight pipe section, the effective overflow area increases and according to the principle that when the fluid flows steadily in a closed pipe, the mass flow rate through any cross-section is equal per unit time.So it produces acceleration and then deceleration at the outlet of the guide vane.Besides, according to the instantaneous flow rate formula, which is calculated as equation ( 1).The calculated value of the outlet velocity of the guide vane pipeline is basically the same as the result of the PIV test value, and the test results are more reliable.
where Q is the volumetric flow rate (m 3 s -1 ), A cross-sectional area of pipeline (m 2 ).Comparing the velocity distributions at different flow rates, there is a significant radial velocity difference in the outlet straight section at 1.1Q des .But, the radial velocity gradient in the straight pipe section of the guide vane outlet at 1.0Q des is smaller and the velocity distribution is more uniform.The flow rate at this point is about twice that of the impeller inlet.

Conclusion
In this paper, PIV is used to measure the flow of the whole basin under different flow rates of the axial flow pump, and the time-averaged velocity field is analyzed to obtain the following conclusions: a.The flow field distribution at the impeller inlet for the three flow conditions is basically the same.The incoming flow follows the contour of the front guide cap to the vicinity of the impeller inlet edge, where the velocity of the flow field gradually increases, forming a high-speed region at the top of the impeller inlet edge.And the acceleration from the hub to the rim is the most pronounced.However, the difference between the highest and lowest speeds at different flow rates is about 1ms -1 .
b.The velocity gradient inside the impeller varies more significantly.The velocity distribution increases in steps from the impeller inlet to the outlet, and deflects towards the outlet of the impeller hub and creates a region of high velocity.As the flow rate increases, the energy gained by the fluid inside the impeller continues to increase, the low-speed zone near the impeller inlet hub is decreasing.At the same time, the high-speed region gradually travelled upwards and the flow field was most uniformly distributed along the axial direction at the rated flow rate.
c.The trend of the velocity distribution of the fluid inside the guide vane is basically the same, the axial velocity of the fluid increases gradually, and the velocity gradient changes significantly.It is important to emphasise that the rectification effect of the guide vanes on the turbulent flow field between static and dynamic is more obvious at the design flow rate.Besides, the fluid velocity distribution is more uniform and the flow is smoother.d.A localised region of high velocity exists in the vicinity of the constriction runner at the exit of the guide vane, and the velocity distribution in the region tends to decrease sharply from the centre of the region to the periphery.In addition, the radial velocity gradient in the straight section of the guide vane outlet at the design flow rate is smaller and the velocity distribution is more uniform.

3. 2 .
The experimental object 3.1.Experimental apparatus The test was carried out on the closed-cycle test bench of the waterjet propulsion pump at the Marine design & Research Institute of China, and the test bench are shown in Figure 2. The test bench includes a model test section, a vertical cavitation cylinder, a flow control valve, an electromagnetic flowmeter, a horizontal steady flow cylinder and other devices.In order to accurately measure the instantaneous speed and torque, a torque speed sensor is connected between the motor and the pump with an accuracy of 0.2%.In the closed-cycle test bench, the flow rate was measured by an electromagnetic flow sensor with an accuracy of 0.2%, and dynamic pressure sensors with an accuracy of 0.1% are installed at a distance of 2D (Pipe diameter) from the inlet and outlet of the model pump, respectively.The flow change in the test bench is achieved by adjusting the coarse control valve DN 600 and the fine control valve DN 150.According to the standard GB/T 18149-2017《Centrifugal, mixed flow and axial pumps -Code for gydraulic performance tests-Precision class》, the test bench meets the requirements of Class 1 accuracy.1. Axial flow pump model 2. Laser device 3. Vertical cavitation cylinder 4. Electromagnetic flowmeter 5. Horizontal steady flow cylinder 6. Flow-control valves Figure Experimental setup.
(a) Impeller inlet shaft section (b) Impeller inner shaft section (c) Guide vane inner shaft section (d) Guide vane outlet shaft section Figure 4. Original figure of tracer particles.

Figure 7 .
PIV measurement results of the guide vane inner shaft section.

Figure 8 .
PIV measurement results of the guide vane outlet shaft section.

Table 1 .
Specifications of the tested pump.