3. RESULTS AND DISCUSSIONS

3. 1 Incident Shock Wave

Prior to discussing the experimental results for cylinder arrays, we describe the basic characteris-tics of the shock tube that is presently used. Figure 7 shows the typical pressure records at port # 1, 4, and 8. There are no cylinders in the test section. The initial pressure in the high pressure chamber (P4) is set as 490 kPa (5 kgf/cm² ) . At each port, almost the same level of pressure rise due to the incident shock wave is observed. The overshoot in the pressure rise at port # 8 is expected to be caused by the slightly recessed (less than 0.5 mm) mount of the pressure gauge in order to protect the sensor from the impact of dusts in the test section. The shock velocity is calculated from the time lag of the pressure rise. The results show that the shock wave propagates at almost constant speed through the test section.

At port # 8 in Fig. 7, the pressure gradually decreases after the plateau behind the shock wave. Such decrease is caused by the expansion wave from the exit of the test section, since the test section is open to the atmosphere and the expansion wave is produced at its end when the shock wave reaches there. The expansion wave propagates upstream in the test section and the time of its arrival determines the effective test duration. The arrival of the contact surface at the test section is much later than that of the reflected expansion wave. In the present experiments, the effective test duration is about 2 msec from the time when the shock wave arrives at port # 1.

Figure 8 shows the variation of the shock Mach number with the initial pressure ratio (P4/P I ) . The experimental shock Mach number is slightly smaller than the one-dimensional inviscid theory ⁶⁾ because of the time required for valve opening and the three-dimensional effect of the flow around the piston valve.

The schlieren picture shows that a plain and undisturbed normal shock wave is generated in the test section. Disturbances due to the change in the cross section shape at the junction between the circular duct and the square duct are not observed. The repeatability of the experimental results is quite good both in the pressure measurement and in the flow visualization.