| Summary: | The need to measure flow rates in fluids is large in the world today. Many are the applications where a more or less accurate measurement is needed, e.g. gasoline pipe, district heating, medical flows and flows in the manufacturing industry. One way to measure flow velocities that is commonly used is with vortex shedding flow meters. These flow meters detect the vortices behind a bluff body in a flow in different ways, one way is by transit time ultrasonics. Since, normally, the sing around-ultrasonic is a more accurate way to measure flow velocity the theory behind this thesis is that a vortex shedder flow meter with sing-around ultrasonics should be more accurate than the ones with transit-time. A vortex flow meter is modeled with FEMLAB to give simulated data for the signal processing, and then the data is processed in MATLAB. The zero-crossing algorithm and the fast Fourier transform are applied to find the vortex frequency. In this stage it seems as if the sampling frequency is too small to detect the vortex frequencies. The final step is to make a vortex flow meter to confirm the theoretical model. The bluff body is built around ultrasonic transducers from D-flow technology AB. The flow meter is inserted to the water laboratory at Luleå University of Technology and measurements are done that is later evaluated and compared to the simulation results. In the real model the vortex frequency is discovered for the lower velocities, but the signal processing used in this thesis is not sufficient to give accurate frequencies for the higher velocities. Two big conclusions are made, that the theoretical model is not good enough and that the vortex frequency is present in the measured signal but could not be found with the algorithms at hand. Validerat; 20101217 (root)
|
|---|