An investigation into the effects of installation on the performance of insertion flowmeters

An electromagnetic and a turbine insertion flowmeter were tested in three different flow conditions inside a 0.590-m bore pipe inserted in the National Engineering Laboratory (NEL) large water flow measurement facility. The results were compared with velocity measurements obtained from a laser Doppler velocimeter (LDV). The advantage of using such a reference measurement is that LDV is non-intrusive and does not affect the velocity profile itself. Of the meters tested, one was supplied with a whole meter calibration factor and the other was supplied with a calibration factor for the D/2 position. For both meters, application of the respective manufacturer's blockage correction improved the velocity measurements, reducing the differences between the LDV and corrected insertion meter measurements and the difference between the integrated insertion meter measurements and the gravimetric measurements. Swirling and skew flow profiles were generated by the installation of the NEL designed swirl generator and flow disturber, respectively. Neither of these disturbed profiles affected the performance of either of the meters in terms of accuracy of measurement compared with the LDV readings. The profiles themselves, however, changed the velocities at the D/8 and 7D/8 points, making single point estimates of the mean velocity inappropriate. A complete 13-point traverse, integrated using the method of cubics as described in BS 1042 [1] (Section 2.3: Measurement of fluid flow in closed conduits, 1992), gave acceptable estimates of mean velocity in both swirling and skew flow for both probes.