CFD Prediction of Hydraulic Effects on External Ultrasonic Flowmeters: A Case Study
External ultrasonic flow meters installed in a nuclear power plant feedwater system were suspected of containing a conservative bias (high reading) despite having been calibrated in a scale model laboratory test. Several independent measures of plant power level indicated that the bias could be as high as 2.0%. The first step in resolving the discrepancy was the construction of a computational fluid dynamics (CFD) model of the hydraulic configurations in which the flow meters were tested in the lab. The CFD model closely matched the original calibration test results. The CFD model for real plant geometry including certain upstream hydraulic features predicted calibration changes of the same sign and order of magnitude as the suspected bias. These upstream elements were absent in the original hydraulic model test. As a consequence of this finding, these additional features were included in new hydraulic calibration tests. Some effects predicted by the CFD model were verified and traceably confirmed, while others were contradicted by the data from the new model. Final calibration test results confirmed the presence of a bias, although not of the magnitude originally suspected. The calibration results provide a traceable measurement basis for use of the flow meters within their specified accuracy after correction for the bias. By use of a chordal ultrasonic flowmeter, velocity profile (pointed-ness) and swirl rates were measured during the tests. The tests, together with the velocity profile information from the CFD simulations, provided insights into the errors and omissions in the original calibration model tests and the assumptions on which the model was based. Finally the calibration tests highlight the strengths and limitations of CFD models for work of this kind.