Abstract
Turbocharger surge remains an area of concern for the automotive industry as it limits the permissible operating range on the compressor map, while also adversely impacting the compressor’s pressure rise, efficiency, and acoustics. The present study uses Stereoscopic Particle Image Velocimetry (SPIV) to investigate the flow field at the inlet of an automotive turbocharger compressor without a recirculating channel. Experiments were carried out at four different speeds, including 80, 100, 120, and 140 krpm, which represent a substantial portion of the compressor map. The mass flow rates investigated ranged from choke to deep surge, thus spanning the entire mass flow regime at each rotational speed. The current work aims to characterize how the compressor inlet velocity field varies with rotational speed, with a specific emphasis on surge. The qualitative nature of the flow field (radial dependence of axial and tangential velocity profiles), over the choke to mild surge range, was observed to be nearly independent of rotational speed for comparable operating conditions (for example, comparison of mild surge at different rotational speeds). A quantitative comparison of the velocity profiles at the choke or mild surge operating points showed an increase in the velocity magnitudes with increasing rotational speed. The flow field at deep surge, however, was observed to change substantially from 80 krpm to 140 krpm. At 80 krpm, the character of the flow field at different times (at different points on the surge cycle) was observed to be similar: the core flow near the center of the duct was always directed into the impeller, whereas the reversed flow occupied an annular region near the periphery in nearly all time instances. However, as the rotational speed was increased to 140 krpm, the variation in the flow field at different instances within a deep surge cycle increased. At 140 krpm, the negative flow rate (where the cross-sectional average flow is directed out of the inducer back into the inlet duct) portion of the surge cycle was still similar to the overall surge flow field at 80 krpm, but over a substantial part of the positive flow rate (cross-sectional average flow is directed into the impeller) portion of the surge cycle, there was no sign of reversed flow within the visualization domain. As the rotational speed was increased, the surge loop (obtained by combining the PIV and pressure transducer data) extended over a wider portion of the compressor map with higher maximum (positive) and minimum (negative) flow rates, along with higher amplitude pressure fluctuations. The mean amplitude of mass flow rate and pressure ratio fluctuations at deep surge increased in nearly a quadratic fashion with rotational speed. The deep surge frequency did not change substantially over the range of rotational speeds examined in this study.
Experimental study on fiber suspensions in a curved expansion duct with particle image velocimetry