| Zugriffsnummer | 46367 |
| Dokumenttyp | Dissertation |
| Peer Review | unbekannt |
| Sprache | Englisch |
| Titel | Simulation based uncertainty analysis of flow rate prediction |
| Autor(in); Institution |
Weissenbrunner, Andreas; 7.5, Wärme und Vakuum, PTB-Berlin
|
| Quelle/Jahr | (2021), XI, 135 S. |
| Dissertationsvermerk | Dissertation, Technische Universität Berlin, 2021 |
| DOI | |
| Verlag | Berlin: |
| Zusammenfassung | Flow meters are usually calibrated in ideal flow conditions. In consequence, flow rate measurements in non-ideal, so-called disturbed conditions are defective. These disturbances are caused by non-straight pipe geometries and are characterized by asymmetrical axial velocity profiles, enhanced turbulence and cross flow components. These flow structures slowly decay in a long straight pipe. In most piping networks elbow configurations are the most common pipe installations, long straight pieces are rarely found. This thesis deals with uncertainty estimation of flow rate prediction in non-ideal flow conditions using numerical methods. Therefore, different measurement principles are applied to simulated flow fields. In contrast to the most other studies of this kind, particular pipe geometries are not studied separately, but rather a class of disturbances is examined by using random inflow conditions from several elbows in a row with variable distances. It is demonstrated that simulation results from a steady state eddy viscosity model are sensitive to their inflow conditions. To analyze the effect of each inflow condition a separate time consuming flow simulation is necessary. Therefore, different effective methods for uncertainty quantification are evaluated. To save additional computation time, a simulation with ideal inflow conditions can be used. To assess the effect of random inflow conditions, the azimuthal orientation of the calculated flow field is set to random. Based on these assumptions uncertainties are quantified downstream several combinations of elbows for • simulated velocity profiles, • velocity-area measurements in a Venturi nozzle, • ultrasonic and electromagnetic flow meters with and without a Venturi/rectangular nozzle as flow conditioner, • single-beam ultrasonic flow meters with different reflecting path configurations, including two suggested new designs, • flow rate prediction with LDV single-path measurements including the LDV-bias error and • a combined laser-optical and numerical method for on-site calibration of flow meters. |
