Abstract

Rake probes are commonly used in turbomachinery applications to measure distorted inlet flows, including both pressure and swirl distortions. An especially common configuration is the 40 probe rake array, consisting of eight identical arms equally spaced circumferentially around the inlet, each arm having five sensing locations spaced along centers of equal area. By measuring many locations simultaneously, rake probes have the advantage of expedited data collection when compared to traversing probes. However, this reduction in test time comes at the cost of rigid geometry with limited measurement locations. As a result, it is possible for rake probes to miss or exaggerate significant areas of the flow profile, such as large gradients or small features, based on the fixed location of the probe and the particular details of the distortion. The purpose of this paper is to demonstrate a procedure that can be used to evaluate the ability of any desired probe configuration (40 probe rake or otherwise) to sufficiently and accurately measure a non-uniform flow profile. Results of this procedure for a range of profiles and probe configurations are also presented.

In order to accurately determine the impact of discrete sampling on the results, two broad sets of data were generated numerically and analyzed. The first set consists of four fundamental total pressure distortions: once-per-rev circumferential, twice-per-rev circumferential, hub radial, and tip radial. The second set consists of three realistic turbofan distortion patterns: two analytic (though not fundamental) profiles, and one generated from S-duct computational results.

For all investigated patterns, Radial Distortion Intensity and Circumferential Distortion Intensity are calculated in the manner described by ARP 1420, a guideline issued by the S-16 Turbine Engine Inlet Flow Distortion Committee for measuring total pressure distortions in turbomachinery. Additionally, interpolated total pressure contours are generated for each measurement configuration. These were then used to make point-to-point comparisons between the actual and estimated data. While total pressure distortion was used as the variable of interest for the majority of this paper, the conclusions may be applied to swirl, temperature, or any other flow property measured using a probe rake or traversing probe.

This content is only available via PDF.
You do not currently have access to this content.