In this paper is presented experimental investigation of the turbulent swirl flow in pipe generated by axial fans. Two various models of industrial axial fans are used. One of these is axial fan W30, model AP 400, Minel, Serbia and has seven blades and outer diameter 0.397m. Second axial fan SP30 is model TGT/2-400-6, S&P, Spain, has six blades and outer diameter 0.386m. This results with greater clearance in the second case. Blades were adjusted for both fans at the angle of 30? at the outer diameter. Test rig length is 27.74-D, where D is average inner diameter app. 0.4 m. Measurements are performed in two measuring sections downstream the axial fans (z/D = 3.35 and z/D = 26.31) with one-component laser Doppler anemometry (LDA) system and stereo particle image velocimetry (SPIV). Obtained Reynolds numbers, calculated on the basis of the average axial velocity (Um) in the first measuring section are for fan SP30 Re = 226757, while for fan W30 Re = 254010. Integral flow parameters are determined such as average circulation and swirl number. Significant downstream axial velocity transformation occurs for both fans, while circumferential velocity is decreased, but non-dimensional velocity profile remains the same. Circumferential velocity distribution for both fans in the central zone corresponds to the solid body, while in r/R > 0.4, where D = 2R, distribution is more uniform. Radial velocity in the case of fan SP30 has almost zero values in the measuring section z/D = 3.35, while its values are significantly increased in the downstream section with the maximum in the vortex core region. On the contrary radial velocity decreases downstream for fan W30 and has also maximum value in the vortex core region for both measuring sections. Level of turbulence, skewness and flatness factors are calculated on the basis of the experimental data. The highest levels of turbulence for circumferential velocity are reached in the vortex core region for both fans. It is shown how statistical moments of the third and fourth order differ from the values for normal Gaussian distribution. In this paper are also analyzed velocity fields by use of SPIV.
Read full abstract