PIV Measurement of Secondary Flow in a Rotating Two-Pass Cooling System With an Improved Sequencer Technique

The flow field characteristics of a two-pass cooling system with an engine-similar layout have been investigated experimentally using the nonintrusive particle image velocimetry (PIV). It consists of a trapezoidal inlet duct, a nearly rectangular outlet duct, and a sharp 180 deg turn. The system has been investigated with smooth and ribbed walls. Ribs are applied on two opposite walls in a symmetric orientation inclined with an angle of 45 deg to the main flow direction. The applied rib layout is well proven and optimized with respect to heat transfer improvement versus pressure drop penalty. The system rotates about an axis orthogonal to its centerline. The configuration was analyzed with the planar two-component PIV technique, which is capable of obtaining complete maps of the instantaneous as well as the averaged flow field even at high levels of turbulence, which are typically found in sharp turns, in ribbed ducts, and, especially, in rotating ducts. In the past, a slip between motor and channel rotation causes additional non-negligible uncertainties during PIV measurements due to an unstable image position. These were caused by the working principle of the standard programmable sequencer unit used in combination with unsteady variations in the rotation speed. Therefore, a new sequencer was developed using FPGA-based hardware and software components from National Instruments (NI), which revealed a significant increase in the stability of the image position. Furthermore, general enhancements of the operability of the PIV system were achieved. The presented investigations of the secondary flow were conducted in stationary and, with the new sequencer technique applied, in rotating mode. Especially in the bend region, vortices with high local turbulence were found. The ribs also change the fluid motion as desired by generating additional vortices impinging the leading edge of the first pass. The flow is turbulent and isothermal; no buoyancy forces are active. The flow was investigated at a Reynolds number of Re=50,000, based on the reference length d (see Fig. 3). The rotation numbers are Ro=0.0 (nonrotating) and 0.1. Engine relevant rotation numbers are in order of 0.1 and higher. A reconstruction of some test rig components, especially the model mounting, has become necessary to reach higher values of the rotational speed compared with previous investigations such as the work of Elfert et al. (2008, “Detailed Flow Investigation Using PIV in a Rotating Sq