Numerical investigation of the error on flow measurements due to exhaust gas heating and cooling in the SBI-configuration

摘要

The possible flow measurement error due to heating or cooling of exhaust gases in the Single-Burning-Item (SBI) test is estimated from numerical experiments. It is illustrated that there is no one-to-one correspondence between the velocity profile shape and the instantaneous Reynolds number, due to the time-dependent temperature and density profile evolution in the exhaust gas pipe. A non-ambiguous relation is found between the velocity profile shape and an 'effective' Reynolds number, based on the turbulent viscosity. Maximum variations of the velocity correction factors, relating the mean velocity to the velocity on the pipe axis, are found to be in the order of 2% for limiting circumstances for the SBI test. The primary effect is caused by instantaneous Reynolds number variations. The effect of heating or cooling of the flow by the hot or cold pipe is noticeable, too. The statements are proved to be valid independent of the computational grid, the turbulence model and the time steps taken to obtain the numerical solutions.