3.1 Vertical Space Distribution Inside Basin
After installing the current measurement jig onto the tank, all the impellers were operated at maximum speed of 592 RPM, and the current flow velocity was measured at the center of the basin (pit center); the results are shown in
Fig. 8. The rotational speed was increased in intervals of 100, 200, 300, 400, 500, 592 RPM. In the initial stage of the graph within 2,000 s, a stairstep graph was observed, whereby the fluid velocity remained flat while the rotational speed RPM was maintained at the same level for approximately 5 min. When the RPM of the current impeller was changed, the fluid velocity converged to a new velocity almost instantaneously. The current velocity stabilized in a short duration. This measurement was performed for more than 4 h at the maximum RPM. After 2,000 s, a constant current velocity was maintained without any significant change. At the maximum RPM, the maximum velocity was measured to be approximately 0.56 m/s in the surface layer.
For frequency analysis,
Fig. 9 shows the spectrum matching of the velocity time series measured in the maximum RPM test shown
Fig. 8 for the 3,000–15,000 s range, where the velocity stabilized and the acceleration and deceleration excluded sufficiently. From each velocity time series, the mean component was removed to conceal the zero-frequency component. In the surface layer (0.1 m), the variation component was small, and a specific frequency component that was relatively larger did not exist. At a depth of 3.1 m, the energy component was large at a frequency of 0.2 rad/s or less. Although no specific resonance frequency existed, a long-cycle vibration component of 30–200 s appeared. This vibration component will not affect the ships and offshore plants that are floating on the surface; however, it will affect slender structures such as mooring lines that are located deep in the water. Therefore, further investigations are necessitated.
During low velocity or deceleration, such as from 200 to 100 RPM, a longer time was required to stabilize the current velocity. Nevertheless, in all cases, sufficient current velocity stabilization was achieved, enabling the model test to be performed within 20 min after reaching the target RPM.
Fig. 10 shows the current velocity measurements at the center of the basin for various depths and RPMs. As mentioned earlier, the surface layer (0.1 m) velocity was approximately 0.56 m/s when all the current pumps were operating at a maximum rotational speed of 592 RPM. This exceeded the initial design target velocity of 0.5 m/s of the current generating system, thereby indicating sufficient performance. For example, in a model test of the offshore plant, which is considered to be the most general, if the model ratio is 1/50, then a model speed of 0.5 m/s would translate to an extremely high velocity of approximately 3.5 m/s on an actual ship. Therefore, all current velocities applied to marine structures can be reproduced.
By interpolating the result at 500 RPM and the velocity at 592 RPM, a 0.5m/s surface layer velocity was expected to be achieved at approximately 530 RPM; therefore, the velocity profile was measured up to a depth of 12.4 m for only 530 RPM. In general, the change in the RPM resulted in a relatively constant change in the velocity. Up to a depth of 4 m, the velocity gradient was almost linear with the depth. The RPM condition was the same for all duct layers; however, a low speed was measured in depths of 5 m and less owing to the large area of the outlet at the fourth, fifth, and sixth layers, i.e., the lower parts of the current generator.
The velocity on the Y- and Z-axes was close to 0, and only extremely small velocities were measured. This implies that the straightness of the flow was obtained in the targeted direction. Although the velocity was less than 0.002 m/s, i.e., significantly lower than the accuracy of the measurement sensor, the sign trend changed at intervals of approximately 1.5 m. Because the three sensors were installed on the measurement jig in 1.5 m intervals, the abovementioned phenomenon was regarded as a consequence of a minute error in the sensor installation angle, not as an actual flow velocity phenomenon. Considering the mounting angle error and measurement error, the mean velocity component on the Y- and Z-axes was non-existent.
Fig. 11 shows the turbulence intensity by depth. The turbulence intensity,
TI, is expressed as shown in
Eq. (1).
Here, σ is the standard deviation of the velocity, and V is the average current velocity. In the surface layer, the turbulence intensity was measured to be 5%–7%. Furthermore, the turbulence intensity increased with a constant slope up to approximately 4 m in depth.
As shown in
Fig. 11, the flow was directed upward from the bottom to the surface. In addition, the inlet was segmented into six layers, and the inlet became narrower toward the upper layers, causing the velocity to be highest in the surface layer. As the flow ejected rapidly from the discharge port and coincided with the free water surface, it dissipated and stabilized. Therefore, almost no spatial velocity gradient was observed in the surface layer, resulting in insignificant velocity fluctuations.
For the lower velocity fluctuations (lower turbulence intensity), it will be difficult to reduce the porosity of the already installed perforated side wall; however, it will be relatively easy to install additional rectifying devices such as a screen at the discharge port. This will enable a constant flow velocity, but the velocity performance will decrease due to pressure loss.
3.2 Horizontal Spatial Distribution of Current Velocity
As previously mentioned, the initially designed target velocity of 0.5 m/s was expected to be achieved by applying a rotational speed of 530 RPM; therefore, the vicinity of the center of the basin was measured at 2 m intervals.
Fig. 12 shows the trend in velocity change as the flow moved from upstream to downstream. In general, a target velocity of approximately 0.5 m/s was measured, with the surface layer velocity being high upstream (near
x = 5 m) and decreasing downstream. In the graph, the position of the origin is the center of the pit, and the direction of the axis is shown in
Fig. 1(b). It is noteworthy that a different trend was shown in deep water. At a depth of 3.1 m, the velocity increased slightly downstream. This is considered to be due to the velocity gradient stabilizing as the high velocity flow in the surface layer became mixed downstream.
A peculiar phenomenon was also observed from the Y-axis. As shown in
Fig. 13, the velocity increased and decreased at regular intervals. This trend was observed in the preliminary CFD analysis at the designing stage, but the effect was excessive in the actual measurement. It was observed that the velocity increased and decreased in 5 m intervals, which is equivalent to the distance between the vertical walls at the discharge port (
Fig. 1(b) shows the vertical wall location). The flow velocity was expected to decrease at the vertical wall due to viscosity; however, a higher velocity was measured. The cause was investigated and shown in
Fig. 14. The velocity decreased at the vertical wall near the discharge port, resulting in a velocity gradient. However, the vorticity in the opposite direction (Z-axis vorticity) to offset this velocity gradient occurred from a distance of 10 m or more from the discharge port, and this rotational component might have increased the flow velocity (i.e., a momentum surfeit).
Other measurement results confirmed that it is advantageous to perform tests slightly downstream from the center of the basin (or pit center) to obtain a more even distribution of the current velocity. However, for some tests that require the 50 m depth of the pit, those tests can only be conducted at the center. In addition, the quality of the waves is expected to be better at the center rather than downstream. Therefore, the appropriate test location must be determined based on the characteristics and purpose of the offshore plant model test, such as the wave quality, current uniformity, and pit utilization. In this study, we obtained sufficient current distribution results from the model test; however, whether more rectifying devices can be added to the discharge port to achieve a more uniform current distribution should be investigated.