The observation was synchronized with the pressure and temperature measurements. A trigger switch attached to the sluice valve started both the FFT and the camera.
3. RESULTS AND DISCUSSION
3.1 EXPERIMENT ON RAPID CONTACT OF SATURATED HIGH-PRESSURE WATER WITH LOW-PRESSURE AIR (SIMULATION OF EDWARD'S PIPE)
For comparison with the experiments in contact between high-pressure saturated water and cold water, experiments were conducted on fast depressurization of a horizontal pipe initially filled with subcooled liquid or rapid contact of saturated high-pressure water with low-pressure air (Edward's pipe). A typical pressure trace obtained in the experiment discharging high-pressure saturated water at 0.9 MPa into a low-pressure air field (atmospheric pressure. 20-25 ℃) is shown in Fig. 2. The sluice valve opening duration was 10 ms. Time 0 ms on the abscissa denotes the time when the sluice valve was fully opened and the trigger switch was on. The saturated high-pressure water and the air initiated contact at time -10 ms. As shown in Fig. 2, the pressure of saturated water decreased from 0.9 MPa to 0.3 MPa in 0.5 ms after the opening. A rarefaction wave generated by the opening caused a pressure decrease from the valve to the closed end of the test section. After the rarefaction wave passed. the pressure increased again and reached a peak at about 10 ms after complete opening. The air pressure oscillated violently for a short period due to flashing of high-pressure saturated water. It took 10 to 20 seconds to balance because air and vapor release from the apparatus to the outside was restricted by the small diameter of the release pipe.
The high-speed video camera could not observe that phenomenon due to the flashing initiated. The phenomenon was simulated by a numerical solution method, NBGS [11-13]. Appendix B explains the analytical model and the settings of boundary and initial conditions. The dotted line in figure2 shows a pressure trace obtained by simulation. A valve opening duration of O seconds was assumed in the simulation, and the air pressure was assumed to be constant at atmospheric. Pressure of the saturated water rapidly decreased from 0.9 MPa to 0.1 MPa within 0.3 ms after contact Although the simulated pressure decreased faster than the experimental result, due to the differing valve opening duration, it closely predicted the experimental result. The results of the experiment and the simulation revealed that no pressure peak occurred in the case of rapid contact of saturated high-pressure water with low-pressure air.
3.2 Experiment on rapid contact of saturated high-pressure water with low-pressure water
Typical pressure traces and video images obtained in the experiments on rapid contact of saturated high-pressure water (1.2 MPa) with low pressure water (0.1 MPa, 23℃) are shown in Fig. 3 and 4. The numbers of the images in Fig. 4 correspond to the time numbers in Fig. 3. In the images, dark areas represent vapor and bright areas represent water. As shown in Fig. 4, a flashing occurred and vapor was generated near the tube wall at 1.3 ms after the full valve opening. A high pressure peak of around 3.5 MPa, which was three times of initial pressure 1.2 Mpa, occurred in saturated water pressure trace PH When the vapor was generated. In low-pressure water or cold water pressure trace PL, this type of high-pressure peak was not identified. This may have occurred because the pressure peak was absorbed in the water-vapor interface and could not be transmitted into the low-pressure water side. The experiments and numerical analysis of saturated high-pressure water discharging into a low-pressure air field showed that no such high-pressure peak was generated.
Fig. 2 Pressure traces in saturated water
The momentum of the low-pressure water might cause the pressure peak in high-pressure water upon contact with low-pressure water in the experiment. Fig.5 shows the relationship between initial pressure of high-pressure saturated water PHO and generated pressure peak PHO. The pressure peak increased with the initial pressure lineally. We have applied a curve fit to collapse the relationship using the method of least squares. The magnitude of pressure peak (MPa) was best arrived at using the following equation;
PHP = 5.02 PHO - 0.88 (1)
As shown in Fig.4, dark area in the image obtained through high-speed video camera represents vapor bubble and bright area water, respectively. An average brightness of test section, where is inner area of dotted circle shown in the figure, was defined as brightness B. Brightness B in the test section thus increases with bubble shrinks and is roughly proportion to volume of liquid phase, i.e., B∝(1-α), where α is a void fraction. In order to confirm the fact of that the pressure peak can be caused by the rapid growth of the flashing vapor bubble generated when high pressure saturated water contacts low pressure water, e.g. a flashing hammer, the relation between the brightness and the pressure was investigated.
