Characterization of a supercharger as boosting & turbo-expansion device in sequential multi-stage systems

Engine downsizing is driving the automotive industry towards more performing, and less polluting drivetrains. However in order to achieve significant engine downsizing without sacrificing performance, these engines need to be heavily boosted. Different boosting configurations can be adopted on an automotive internal combustion engine, e.g. using turbocharger and supercharger either separately or in series or in parallel. However for highly downsized engines, multistage boosting systems configurations are becoming unavoidable due to the high charge pressure required on the engine intake. Besides the technical challenges associated with achieving adequate tuning, interoperability and driveability of multistage boosting systems, another challenge lies in their performance prediction during engine design. Indeed, performance maps of single boosting systems are usually provided by manufacturers and used as look-up tables in 1-D engine models. Tests are usually conducted in a standalone mode, with no information provided on the behaviour and performance of the combination of more than one system. Aim of this work is to provide an insight into the experimental characterization of a high-pressure supercharger in a sequential boosting arrangement.
The UltraBoost project aims to achieve extreme engine downsizing by means of a turbo-supercharger arrangement for a 2.0L-4 cylinders gasoline capable to deliver the same output torque as an equivalent naturally aspirated 5.0L-V8 baseline (i.e. 60% engine downsizing). The aim of this paper is to assess the performance of the supercharger under different operating conditions. A dedicated test facility was set-up at Imperial College London to evaluate the supercharger performance under different inlet conditions: firstly an initial set of tests was performed with pressure and temperature at ambient conditions in order to calibrate and validate the test facility against the manufacturer performance maps; secondly the supercharger was tested with varying inlet pressures and temperatures matching on-engine operating conditions and the results were then used to assess the effectiveness of 1-D engine models performance prediction when dealing with multistage boosting systems. In addition to this, an assessment on heat transfer in superchargers was also carried out together with the analysis on the nature of non-dimensional performance maps when dealing with a pressurized inlet. Finally, the analysis also looked into the opportunity to use the superchargers as expanders (‘expansion mode’) in order to cool the air charge entering the engine; experiments were carried out.
The results showed that there is discrepancy between the efficiency values computed by 1-D engine models and those obtained experimentally under pressurized/heated inlet air conditions; the correction of the efficiency maps for heat transfer plays a significant role in the final measured efficiency and the correction of the maps for varying inlet temperatures must be carried out in order to avoid incurring in apparent efficiencies greater than unity. The experiments on the supercharger in ‘expansion mode’ showed that low isentropic efficiencies can be achieved; despite this, preliminary simulations on the UltraBoost engine showed that it is possible to achieve savings of few percentage points in BSFC when the supercharger is used to recover some throttling energy by expanding the close-to-ambient pressure to the required intake pressure.