Experimental determination of the dynamic response of Coriolis mass flow meters

The dynamic response of Coriolis meters is significant in many applications, including fast control operations, e.g. short duration (seconds or less) batch-filling, closing flows, and the potential for application to static gas turbine engine fuel flow control. The dynamic response of a meter is determined firstly by the dynamic response of the flow tube (as detected by the two motion sensors) and subsequently by the data sampling and signal processing algorithms used to extract the phase-difference to generate the user output. The flow tube dynamic response and meter indicated response (pulse output) were determined experimentally for a number of commercially available meters, by subjecting each meter to step changes in flow rate. The fastest steps achieved were of duration 4.5 ms. It has previously been shown that the meter flow tube response time, as extracted through phase-difference measurements, cannot be less than the duration of one drive cycle of the tube vibration. Correspondingly, flow tube dynamic response times in the range of 1.4-10 ms were observed (for meter drive frequencies (approximate) in the range of 700-100 Hz). As predicted by theory (straight tube) and finite element simulation, the flow tube step response also includes contaminating (noise) components associated with the Coriolis frequency. There are indications that this noise amplitude was increased by mechanical vibration effects induced by the flow step mechanism. As expected, the meter user output (pulse) indicated much slower step responses than those of their respective flow tubes. These outputs were characterised by a delay in the onset of the step and subsequent lengthening of the step duration which was associated with the output update rate. In some cases, the step noise was apparently eliminated in the user output and this effect was enhanced by the relatively slow update rate. (c) 2005 Elsevier Ltd. All rights reserved.