Simulation of Airflow Through a Test Chamber

Army Research Laboratory

Wednesday, April 01 2009

An apparent anomaly was resolved and a correction was devised.

A computational-simulation study of the flow of air through a thermo-anemometer chamber was performed to resolve what originally seemed to be an anomaly in the measurement data obtained by use of the chamber. The thermo-anemometer chamber is a test chamber used to measure the rate of generation of heat by a device placed within it. In the original application that produced the apparent anomaly that prompted this study, the chamber was used to measure the power dissipation (as manifested by heating) in an operating power-supply inductor. The apparent anomaly was that the heating of the inductor as calculated from the measurements made by use of the chamber seemed unrealistically high.

The thermo-anemometer chamber (see Figure 1) includes a thermally insulating box with inlet and outlet holes. A blower at the inlet forces air through the box. There are a thermo-anemometer and a thermometer at the outlet and a thermometer at the inlet. In principle, the rate of generation of heat by a power-supply inductor or other device in the chamber can be calculated from the outlet area, the outlet air speed (as measured by the thermoanemometer), the barometric pressure, the relative humidity, the inlet and outlet air temperatures, and the specific heat of air and water vapor. For the purpose of computational simulation, the chamber is deemed to be also equipped with an outlet tube that serves to average the flow somewhat and to help suppress vortices, which could be problematic for interpretation of simulation data.

The simulation of airflow in the chamber was performed by use of a computational fluid dynamics program called COSMOSFlowWorks. As a compromise between avoiding computational anomalies (necessitating a fine mesh) and avoiding excessive computation time (necessitating a coarse mesh), the chamber was divided into a computational mesh of 48,000 cells. The inlet flow speed was assumed to be constant at 5.1 m/s. Computations were performed for a series of outlet diameters ranging from 2 to 4" (5.08 to 10.16 cm).

The simulation results revealed the source of the apparent anomaly to be a combination of Bernoulli-like and laminar-flow effects at the outlet. For example, in Figure 2, which depicts results for the case in which the inlet and outlet diameters are both 3" (7.62 cm), the axial velocity at the edge of the outlet is reduced and the axial velocity in the center of the outlet exceeds the inlet velocity; the effect on the axial velocity averaged across the outlet area is equivalent to that of reducing the outlet diameter. This effect causes the thermoanemometer, located in the center of the outlet, to read a speed greater than the cross-sectional average, so that the use of this speed in estimating the heat-dissipation rate gives rise to an unrealistically high value. The simulation results make it possible to calculate an effective reduced diameter of the outlet port to correct the apparent anomaly.

This work was done by Gregory K. Ovrebo of the Army Research Laboratory.
ARL-0029