4.1. Validation of the Numerical Analysis Results
To validate the unsteady-state numerical analysis results of the Francis turbine model, this study compared the results of the steady and unsteady-state RANS equations with the experimental results of the full-scale Francis turbine, as shown in
Figure 5. The efficiencies were normalized by the maximum value of experimental efficiency. In the unsteady-state numerical analyses, two low flow rate conditions with GV angles at 16° and 12° were selected as the observed low flow rate conditions based on the best efficiency point (BEP) with a GV at 22°. The full-scale Francis turbine investigations were conducted using the pressure–time method with a measurement error of ±1.74% [
23]. To compare the efficiencies between the model and the full-scale Francis turbine, the scale-up conversion of the hydraulic efficiency defined by IEC Standard 60193 was applied to the results of the Francis turbine model’s analysis for both steady and unsteady-state RANS [
14]. Equations (7)–(9) were applied as the formulae for scaling up the hydraulic efficiency, whereas Equation (7) was considered as the loss efficiency due to the model’s geometrical scale. The equations of loss efficiency were calculated as a function of the Reynolds number along with Equations (8) and (9), as follows:
where the subscripts of M and P indicate the model and the full-scale Francis turbine, respectively, and
= 7 × 10
6,
= 0.7 are the reference values of a radial turbine defined by IEC Standard 60193 [
14]. The results of the unsteady-state numerical analyses were averaged during the last three revolutions of the runner to enable an efficiency comparison.
The comparisons between the experimental and the conversed numerical results demonstrate a slight variation for each efficiency. However, the tendencies of the conversed efficiencies of the numerical analyses were similar to those exhibited by the experimental efficiencies; these variations in efficiency comparisons can be interpreted as the neglect of the mechanical loss and the surface roughness in the numerical analyses. Consequently, this study considered the numerical analysis results for the Francis turbine model to be valid. However, the addition of anti-cavity fins decreased the efficiencies across the entire range of observed flow rates. Particularly, in the unsteady-state analyses, the efficiencies of the GVs at 22°, 16°, and 12° decreased by 0.5%, 0.8%, and 1.0%, respectively, depending on the anti-cavity fins applied.
4.2. Internal Flow Characteristics Relative to the Anti-Cavity Fins in the Draft Tube
This study investigated the performance characteristics concerning the anti-cavity fins in the DT and the flow rate conditions by calculating the head losses of the main components of the Francis turbine model, as presented in
Figure 6. The model’s head losses were calculated using Equation (10) for the SC, SV, GV, and DT, whereas Equation (11) was applied to calculate the head loss of the runner [
24].
Here, Hloss represents the value of loss by a head, ∆ptotal is the total pressure difference through each component, which was calculated with time-averaged total pressure in this study, ρ is the water density, and g is the acceleration due to gravity. In the Hloss runner, T is the torque of the runner, which was measured by the force caused by a rotating axis, and ω is the angular velocity.
Similar head loss distributions were exhibited from the SC to the GV regions both with and without anti-cavity fins, whereas the inclusion of anti-cavity fins caused slight decreases in the runner’s head losses. However, the application of anti-cavity fins in the DT increased the head losses; these losses were comparatively greater than the decreases in head loss in the runner region. In particular, the difference in head loss in the DT region due to the addition of anti-cavity fins was about 2.0% with a GV at 16°, which was relatively the highest under observed conditions of flow rate.
To investigate the qualitative effect of anti-cavity fins to the magnitude of vortex rope in DT according to the flow rate conditions,
Figure 7,
Figure 8 and
Figure 9 show the vortex rope in the DT by the iso-surface distributions of pressure during the last revolution of the runner for investigating the internal flow structures of GVs at 22°, 16°, and 12°. The iso-surface of pressure was determined as the relative water saturation pressure considering the water level of the lower reservoir. As
Figure 7b reveals, with a GV at 22°, the vortex rope in the low-pressure region was not generated in the DT without anti-cavity fins. However, as can be seen in
Figure 7a, the application of anti-cavity fins produced low-pressure regions in the DT. Due to the protrusion of the air injection outlets on the anti-cavity fins, the low-pressure regions occurred on the anti-cavity fins near the inlet of the DT (as detailed in
Figure 1). Therefore, it is believed that inducing flow resistance at the sites of the air injection outlets generated the low-pressure regions.
Figure 8 shows that the vortex rope was clearly developed in the DT with a GV at 16° (0.78 Q
BEP). The application of anti-cavity fins in the DT significantly decreased the vertical length of the vortex rope. Additionally, the low-pressure regions were found to occur near the anti-cavity fins.
Figure 9 shows the condition of the GV at 12° (0.59 Q
BEP). Here, and similar to that shown in
Figure 7a, without the addition of anti-cavity fins, no low-pressure regions in the DT were generated. However,
Figure 9a reveals that the low-pressure regions were developed on the anti-cavity fins rather than those with a GV at 22° as the difference under both conditions of GV. In this way, the anti-cavity fins in the DT show the effect of reducing the vertical length of the vortex rope, but the shape itself, such as in the air injection outlets, acts as the factor that induces the low-pressure regions and impedes the flow in the DT. It can be regarded as the cause of the increased head loss by application of the anti-cavity fins in the DT, as shown in
Figure 6.
To observe the flow phenomena in the DT according to the flow rate, this study compared the velocity triangles at the runner outlet with GVs at 22°, 16°, and 12°, as illustrated in
Figure 10. As the meridional velocity, C
m, decreased, the absolute flow angle, α
2, gradually increased as the GV angle decreased. Thus, in the absolute velocity component, C
2, the radial velocity component increased as the α
2 also increased. The increase in the swirl strength of the flow and the generation of both the complicated flow and the vortex rope in the DT can be due to the increase in the radial velocity component at the outlet of the runner under conditions of low flow rate.
The streamline distributions of the time-averaged (Trnavg) velocity in the DT both with and without the anti-cavity fins are presented in
Figure 11 for GVs at 22°, 16°, and 12°. The flow phenomena in the DT are shown with the associated complicated flow at low flow rates as confirmed by the flow characteristics at the runner outlet in the velocity triangles in
Figure 10.
Figure 11b presents the complicated flow with precession in the DT with a GV at 16°. As
Figure 8 shows, the internal flow makes it possible to confirm the cause of the development of the vortex rope and the low-pressure region near the anti-cavity fins in the DT.
Figure 11c presents a very complicated internal flow pattern without precession due to the increased radial velocity component with a GV at 12°. These flow characteristics can be confirmed as the reason for the development of the low-pressure regions on the anti-cavity fins, as shown in
Figure 9a.
The angle distributions of the absolute and relative flows at the runner outlet along the spanwise direction from hub (0) to shroud (1) with GVs at 22°, 16°, and 12° are presented in
Figure 12. The observed flow angles are normalized by the value of each maximum flow angle. The GVs at 22° and 16° demonstrated similar distributions of flow angles both with and without anti-cavity fins, whereas the GV at 12° produced a slight difference in flow angle distributions. This is due to the complicated internal flow characteristics induced by low flow rate conditions. The absolute flow angle distributions at the 22° GV are close to 0, and the absolute flow angles increase as the flow rates decrease. Additionally, compared with the other GV conditions, the contrasting trends are exhibited only at the GV at 12° near the spanwise range of 0–0.2. Similarly, the GV at 12° also demonstrates different tendency characteristics in relative flow angle distribution due to the complicated internal flow, as shown in
Figure 12b, although the relative flow angle and the angle of the runner blade outlet generally appear similar.
The above analyses reveal that at the observed GV of 16°, the head loss difference in the DT was the greatest, and the effects of the anti-cavity fins were significantly observed.
Figure 13 presents a more detailed analysis of the internal flow characteristics of the same GV; it investigates the influence of the anti-cavity fins by comparing the iso-surface distribution of the pressure in the DT during the last revolution of the runner with the top view of the runner. The pressure value was determined to be the same as that used in
Figure 8. Depending on time and irrespective of the inclusion of anti-cavity fins, the vortex ropes were maintained at similar maximum rotating diameters. Therefore, the effects of the anti-cavity fins decreased the maximum vertical length of the vortex rope, whereas the maximum rotating diameters were not significantly affected by the inclusion of anti-cavity fins.
To investigate the effect of the anti-cavity fins in detail,
Figure 14 and
Figure 15 present the pressure distributions on the observed cross-sections (as shown in
Figure 1) at the GV at 16° with and without anti-cavity fins, respectively. The cross-sections were examined at 0.1, 0.3, 0.6, and 0.9 D
2 in the direction of flow from the inlet of the DT, and the pressure values were normalized by the value of maximum pressure. The low-pressure regions in each cross-section of the pressure distributions were maintained with similar diameters, depending on time, and there was a tendency for the low-pressure regions to gradually decrease along the direction of flow. Furthermore,
Figure 14 indicates that the low-pressure regions were clearly developed up to the cross-section of 0.3 D
2; however, because of the influence of the anti-cavity fins, these regions decreased considerably from 0.6 D
2. Meanwhile, the low-pressure regions were formed near the anti-cavity fins. It can be regarded that these regions were induced by the resistance of the anti-cavity fins to the tangential component of the absolute velocity, which increased as flow rate decreased as shown in
Figure 10 (velocity triangle). Thus, via the low-pressure regions generated near the anti-cavity fins in
Figure 14b, the formation sites of the low-pressure regions in
Figure 8 and
Figure 13 can be confirmed.
Figure 15 clearly shows the low-pressure regions in the DT up to the cross-section of 0.9 D
2 without anti-cavity fins. Actually, the existing DT plays a role of recovering the static pressure in the flow; however in the low flow rate condition, the low-pressure regions indicated by the vortex rope with precession decreased with the anti-cavity fins rather than the decrease through the role of the DT itself. Therefore, it can be concluded that the anti-cavity fins effectively suppressed the generation of the vortex rope in the DT, particularly by decreasing the vertical length rather than by modifying the rotating diameters.
To analyze the influence of the anti-cavity fins at GVs at 22°, 16°, and 12°, the time-averaged axial and circumferential velocity components were compared on the observed line of 0.6 D
2 (in
Figure 1) in the DT, as shown in
Figure 16. The abscissa indicates the measurement location relative to the diameter from the wall (0) to the wall (1) of the DT, and the velocity values were normalized by the value of maximum velocity. Without the addition of anti-cavity fins, the axial velocity of the GV at 22°, as shown in
Figure 16a, decreased slightly near to the wall, whereas the relatively small velocity range altered according to the anti-cavity fins.
However, the overall greatest difference in the axial velocity distribution was shown by the GV at 16° relative to the anti-cavity fins, and the backflow occurred near d/D = 0.5. With the GV at 12°, the backflow was generated in a relatively wide range of d/D = 0.2–0.8. A complicated internal flow without a vortex rope was demonstrated, and the difference in axial velocity was not shown to vary significantly according to the use of anti-cavity fins. Therefore, the anti-cavity fins had a relatively significant effect on the axial velocity at the GV at 16°; here, a vortex rope was formed, which was due to the shape characteristics of the anti-cavity fins installed in the axial direction concerning the flow direction.
In terms of circumferential velocity distributions, the anti-cavity fins near the wall of the DT revealed a slight difference at the GV at 22°, as presented in
Figure 16b. The difference between the maximum and the minimum circumferential velocity was relatively greater at the GV at 16° without the anti-cavity fins; the addition of anti-cavity fins effectively reduced this difference. Furthermore, because of the complicated internal flow, a difference in the circumferential velocity distribution was exhibited along with the anti-cavity fins at the GV at 12°. Therefore, this study considers that the anti-cavity fins influenced the velocity component in the circumferential direction rather than the axial direction in the DT; the internal flow was mainly influenced under the low flow rate conditions, during which the vortex rope was generated.
4.3. Unsteady Pressure Characteristics Relative to the Anti-Cavity Fins in the Draft Tube
To investigate the unsteady pressure characteristics according to the use of anti-cavity fins in the DT, this study compared the unsteady pressures obtained via fast Fourier transformation (FFT) analysis with GVs at 22°, 16°, and 12°, as shown in
Figure 17. The pressure measuring points of 0.1, 0.3, 0.6, and 0.9 D
2 were used in the flow direction on the DT wall, and the value for the highest pressure was used for each height from the four measuring points (p1–p4), as indicated in
Figure 4. The maximum magnitude normalized the pressure values, and the frequency was normalized by the rotational frequency, f
n, of the Francis turbine model. The highest first blade passing frequency (BPF) was shown on the 0.1 D
2 at the GV at 22° as the BEP condition. However, at the normalized frequency of 0.37 f
n in the low-frequency region, relatively high-pressure characteristics were demonstrated before the first BPF at the GV at 16°, as the vortex rope with precession developed compared with other GV conditions. Previously, Kim et al. [
25] numerically investigated similar unsteady pressure phenomena in the low-frequency region due to the vortex rope in the DT. Furthermore, the addition of anti-cavity fins at a measuring height of 0.1 D
2 slightly increased the unsteady pressure at the GV at 16°. This was due to the effect of the anti-cavity fins in reducing the length of the vortex rope of the GV at 16°, as can be seen in
Figure 8,
Figure 14 and
Figure 15. However, the vortex rope remained near the inlet of the DT.
Unsteady pressure characteristics were exhibited at the normalized frequency of 0.37 f
n and near the low-frequency regions before the first BPF at the GV at 12°, where a complicated internal flow without a vortex rope was evident. At all the observed GVs, the first BPF was gradually decreased from 0.3 D
2 (shown in
Figure 17b) in the flow direction. Without the addition of anti-cavity fins, the GV at 16° at 0.3 D
2 demonstrated greater unsteady pressure characteristics at the normalized frequency of 0.25 f
n than it did with them. Unsteady pressure at the 12° GV at 0.3 D
2 increased in the normalized frequency range of 0.5–0.75 f
n, which was greater than the magnitude at 0.1 D
2. In
Figure 17c, both unsteady pressures at the GV at 16° increased more at the normalized frequency of 0.25–0.37 f
n compared with those at 0.3 D
2; at the 12° GV, the normalized frequency of 0.75 f
n was shown to be similar to the magnitude of 0.3 D
2. As is evident in
Figure 17d, the GV at 12° exhibited similar unsteady pressure characteristics to those at 0.6 D
2.
The differences in the maximum magnitude of unsteady pressure due to anti-cavity fins were noticeable in the GV at 16°, and the magnitude of unsteady pressure with the anti-cavity fins was similar to the magnitude with anti-cavity fins at 0.6 D2. Therefore, the unsteady pressure characteristics increased along the flow direction in the DT under conditions of low flow rate with a developed vortex rope; the use of anti-cavity fins effectively decreased the unsteady pressures. Particularly, the application of anti-cavity fins resulted in an approximate 41% reduction in the maximum magnitude of unsteady pressure at the 16° GV at a measured height of 0.9 D2.