Velocity Vector Determination from Multiple-Sensor Pneumatic Probe Measurements

Abstract

Theme D data for compressors are often derived from measurements made with multiple-sensor pneumatic probes. Although it is generally admitted that the presence of a probe alters the flow to some extent, few attempts have been made to evaluate and to correct probe readings for the interference that they produce. The goal of the present work was to measure the velocity distribution between blade rows of a small transonic compressor (11-in. o.d. and 2-in. blade height), operating subsonically. Different probes were used in the compressor annulus and were also applied to measure uniform temperature flows in pipes, free jets, and annular ducts. A useful method for representing probe characteristics and correcting for the presence of flow boundaries was devised. The probe calibration away from boundaries was represented by two interrelated polynomials for the pitch angle and velocity magnitude. In application, the probepressure measurements were reduced, using a computer offline and distribution of temperature (measured separately) to velocity magnitude and pitch angle. The departure of probe measurements near a flow boundary was then treated as being a self-induced flow effect. The effect was examined in known flows and represented analytically. The analytic representation of the calibration and of the corrections allowed a routine reduction of probe data from compressor surveys. The method can be applied generally to probes having four or five sensors for the measurement of pitch angle, yaw angle, and velocity magnitude. The measurements made with different probes near to flow boundaries showed that errors could be surprisingly large or very small depending on particular design features. A new probe which also incorporated a temperature sensor was designed and used successfully to transonic Mach numbers.