PIV for Turbomachinery Applications

Author: Mark P. Wernet

Turbomachines are used in a wide variety of engineering applications for power generation, pumping and aeropropulsion. The need to reduce acquisition and operating costs of aeropropulsion systems drives the effort to improve propulsion system performance. Improving the efficiency in turbomachines requires understanding the flow phenomena occurring within rotating machinery. Detailed investigation of flow fields within rotating machinery have been performed using Laser Doppler Velocimetry (LDV) for the last 25 years. LDV measurements are time and ensemble averaged over all of the blade passages in a rotating machine. PIV measurements enable the study of non-stationary flow phenomena, and hence is a very powerful diagnostic. Moreover, a series of instantaneous spatial velocity measurements obtained using PIV can be averaged together to compute the mean flow field for comparison to LDV and CFD results.

Recent advances are leading to the emergence of PIV as a powerful velocity measurement technique which can be used as an alternative or complementary approach to LDV in a wide range of research applications. PIV is a planar measurement technique wherein a pulsed laser light sheet is used to illuminate a flow field seeded with tracer particles small enough to accurately follow the flow. The positions of the particles are recorded on either photographic film or digital CCD cameras at each instant the light sheet is pulsed. The use of CCD camera systems employing the "frame straddling" technique developed by Wernet in 1991 enable electronic capture of single exposure PIV images with very short (<1 microsecond) inter-exposure times. These single exposure image pairs are then processed using cross-correlation techniques yielding wide dynamic range and directionally resolved velocity estimates. Refined data processing techniques and continuous increases in computational power have made PIV a more widely available and practical measurement technique. Near real time display of the velocity vector maps computed from the acquired image data are now possible.

Due to the high rotational speeds typically encountered in aerodynamic turbomachinery and the concomitant high flow velocities, pulsed Nd:YAG lasers are required to provide sufficient light energy (~100mJ/pulse) in a short time interval (<10 nsec) to record an unblurred image of the particles entrained in the flow. Backscatter LDV systems are reasonably well suited for the single optical access port typically encountered in turbomachinery applications. However, the standard PIV technique requires that the light scattered by the particles traversing the light sheet be collected at 90 degrees from the plane of the light sheet. Hence, the light sheet must be introduced either upstream or downstream from the optical access port used for scattered light collection and directed along the flow direction. Care must be taken to insure that the light sheet optics do not disturb the stream tube feeding the blade passage under study. If the propagation direction of the light sheet is aligned with the stagger angle of the blades and the optics are located sufficiently far upstream, then the flow in the measurement region will not be disturbed.

A very compact light sheet delivery system was constructed using a periscope type configuration. The pulsed Nd:YAG beam is directed down the bore of the tube which contains light sheet forming optics and a 90 degree turning mirror. The light sheet exits the probe through a window which keeps the optics inside the probe protected from contamination by the seed material. An implementation of this NASA designed and constructed probe is shown in operation in figure 4. The small diameter periscope probe is inserted through the compressor casing upstream of the measurement location, see figure 5. Moving the probe in and out through the casing changes the spanwise location of the illumination plane.

In the periscope probe configuration, directing the pulsed laser beam down the bore of the periscope tube is extremely challenging. An articulated light arm with knuckled mirror joints was used to easily and reliably direct the beam down the bore of the probe. Articulated light arms for Nd:YAG laser beam delivery are commercially available from several commercial PIV vendors and permit the full energy range of the laser to be used (200 mJ/pulse at 532 nm). Use of the light arm simplifies the coupling of the Nd:YAG beam to the periscope and also adds an increased level of safety to the installation since the beam is entirely enclosed when outside of the compressor casing.

The world's first ever successful digital PIV measurements in a transonic compressor have been demonstrated at NASA Glenn. Measurements were obtained in a single stage 50.8 cm diameter transonic axial compressor facility. The measurements were obtained using upstream illumination of the blade-to-blade rotor passage at a rotational speed of 17,150 rpm. A sample velocity vector map obtained by averaging 110 processed velocity vector maps together is shown in figure 6. The blade profiles at the measurement plane location are shown in the figure. The flow is from left to right and the rotational direction of the rotor is from top to bottom. Under these rig operating conditions a strong shock forms off the upper blade leading edge and spans the blade passage. The position of the blade-to-blade plane shock is readily observed by the sharp drop in vector magnitude within the blade passage (pink to blue shading). A bow wave also forms off of each blade extending outward (up and to the left) from each blade leading edge. The bow wave from both the lower and upper blades are observed in the left portion of the image. There is also a significant change in velocity magnitude across the bow waves as indicated by the color shading (yellow vectors). The total acquisition time for the data shown here was approximately 6 minutes. The averaged measurements illustrate that PIV yields high accuracy velocity vector maps in much less time than traditional LDV techniques.

Figure 4  Figure 5  Figure 6