15th European Conference on Turbomachinery Fluid dynamics & Thermodynamics
Paper ID:
ETC2023-165
Main Topic:
Experimenta Techniques
Authors
Abstract
Vast progress has been achieved in the last decades on understanding the phenomena related to the onset of condensation in steam flows, both experimentally and especially numerically. Nevertheless, there is still a considerable disagreement between the various numerical models used. Unfortunately, the available experimental validation data is not sufficiently detailed to allow for a proper validation of CFD simulations. Hence, further experiments are necessary to close this gap. This paper presents new experimental data of a condensing steam flow, acquired in a supersonic nozzle according to Barschdorff, at the Institute of Thermal Turbomachinery Laboratory (ITSM) in the University of Stuttgart. In previous experiments done at the ITSM, steam was throttled and cooled down using a controlled water injection. However, the stability of inlet conditions, in particular the inlet temperature, was not satisfactory. Moreover, evidence of larger droplets was detected during the post-processing of data. Consequently, this paper also introduces novel modifications to the ITSM nozzle test-rig, which consist of adding a water bath steam cooling vessel system to accomplish steady inlet conditions and improve control of the operating points in the nozzle. With this scheme it is possible to maintain stable and dry inlet conditions, yielding in having sufficient time to measure the flow features and droplet spectra repeatedly, generating new and reliable experimental data. Similar to the experiments done by Barschdorff, a steady inlet pressure of 784 mbar was set at various inlet temperatures down to 100.2°C. Condensation onset is accurately captured across the nozzle for all operating conditions using up to 1 mm spatial resolution for both pneumatic and droplet size measurements. Droplet light spectra are measured using the light extinction method. CFD simulations were performed using the commercial solver ANSYS CFX; Sauter diameters, droplet numbers, wetness and its correspondent condensation onset locations were calculated. Nonetheless, the droplet diameters are numerically overestimated by approximately a factor of 2. Thus, it is evident that further understanding and development of the numerical droplet growth model, especially at lower expansion rate nozzles, is still obligatory. The validity and accuracy of this work is demonstrated by the extremely high reproducibility of the results. It is concluded that there is a good agreement pneumatically, between experimental and numerical results.