Aerodynamic drag of a modern car is generated mainly by underbody flows.A better understanding of these flows and of their interactions with the car underbody, may contribute to the future improvement of the car drag characteristics.This paper reports the results of a parametric study carried out in the Pininfarina wind tunnel, on a full scale simplified car model, by using the Ground Effect Simulation System built in 1995.The main aim of this study was to investigate the effect on the aerodynamic coefficients produced by important geometric changes which affect the flows under the car, in proximity of the ground, and are often difficult or impossible to be modified when tests are made on real cars.The model chosen for this research program is that defined by the SAE “Open Jet Interference Committee” as a reference model to be used for investigating wind tunnel interference and for comparison between wind tunnels. In particular it has no wheels.The model built in full scale by Pininfarina is, in reality, a super-set of the SAE model, as it includes a number of additional parts, which are necessary to investigate the behavior of this model in various conditions of ground effect.These add-on parts include: Front and rear wheel-housings of three different sizes. Wheels of 2 different sizes / covers Some different wheel tracks Three different front and rear overhangs. Three different rear diffusers. These parts have been interchanged, to cover a great number (not all) of the possible different configurations.In addition, each of these configurations has been tested: At various ground clearances, from 200 to 50 mm In the four classical GESS conditions, namely Moving Ground and Rotating Wheels Moving Ground Only Rotating Wheels Only All Static In addition, pressures have been measured along the underbody centerline, including the diffuser and the model base.Furthermore the flow field under the model has been measured for some selected configurations, to achieve a better understanding of the data obtained by the balance.The analysis of all this large number of measurements has been rather engaging.A multiple regression analysis has been used to estimate the change which can be expected for each coefficient, CD, CL, CLF, CLR, CMY, when the model configuration and /or its ground clearance is changed.These results are reported in the paper, together with some selected surface pressure diagrams.It is worth noting that this is the first time that the results of a systematic study into ground effect at full scale, have been published.Although these results are specific for this model shape, they can contribute to a better understanding of the effect of the underbody layout changes as well as of the differences which can be expected for the various GESS testing conditions.It is well known that the largest component of car aerodynamic drag arises from the energy losses in the flow along the car underbody, the flow within the wheel-housings, and the flow around or through the wheels. Evidence of that was shown by flow maps published several years ago (1) (2). One of these maps, reprinted in fig.1, shows clearly that most of the momentum loss at the back of the car, i.e. the most important contribution to the car CD, is located at a height over the ground between 0 and 500 mm. The car shown in the map, although designed about ten years ago, had a remarkably low drag coefficient ( CD ∼ 0.29) , not too different from the CD values of the best cars which are sold nowadays. Therefore the flow conditions shown behind this car are still representative of those of many present passenger cars.This also means that any realistic possibility of further significant improvement to the aerodynamic drag of a modern car is strongly dependent on reducing the losses from the underbody flows.However, working on improving underbody flows has been historically difficult, for a number of reasons.Car manufacturers are reluctant to invest money (or too much money ) for streamlining the car underbody. The Opel Calibra (3) launched in 1986 is probably the first example of a mass production car where underbody aerodynamics has been taken into account (fig. 2), and that has certainly contributed to the excellent CD value of this car (0.26 - 0.28, depending on the wind tunnel).The reluctance of car manufacturers to accept aerodynamic development of the car underbody is due to several factors, namely the increase of cost and weight associated with fitting the shields to the underbody, the cost of redesigning some parts of the underbody, the possibility of thermal problems, and eventually a more difficult maintenance of the drive line.Beside these reasons, which are certainly partly true, an additional important reason which discouraged aerodynamic work on car underbodies was the lack of experimental facilities able to reproduce flow conditions under the car which are sufficiently similar to the conditions existing on the road.In reality the possibility of reproducing in full scale the car to ground relative motion and, at the same time, the spinning of the car wheels, is relatively recent. DNW reported tests of this type on passenger cars in 1989 (4), and FIAT in 1990 (5) (6). Both used (and still use) a large belt, and that is certainly an excellent way to provide a good simulation of the road condition. However the system to support the car over the belt is relatively complex in both wind tunnels, and require lengthy preparation of the car and /or the wind tunnel. This last point has somehow discouraged the use of these large moving grounds for every day development work.More recently, in 1995, Pininfarina has built in its Wind Tunnel a Ground Effect Simulation System (GESS) of new design. It includes various sub-systems, namely a narrow moving belt (MB) to cover the area under the car between the wheels, separate rotating rollers (RR) to spin the wheels, a boundary layer suction system (BLSS) followed by a tangential blowing system (TBS), both upstream of the belt. The latter feeds the belt with a correct boundary layer and reduces as much as possible the boundary layer displacement thickness on that part of the test section floor not covered by the moving belt.Details on the GESS and on its latest upgrades are reported in (7) (8) (9). Fig. 3 shows the various sub-systems and in particular the new rollers.The main advantage offered by this new type of GESS is its high level of practicality. Almost any passenger car can be tested in conditions of moving ground and rotating wheels, with a very short preparation. And that has made possible and practical the aerodynamic development of the car underbody.The research program which is reported in this paper is the logical follow-up to the availability of this Ground Effect Simulation System, and demonstrates the possibility of performing a large number of tests, easily and with a good simulation of the underbody flows.Beside that, the following points have been taken in account. 1. In three years of operation the GESS has been used to test and/or develop more than 100 cars having different body shapes, underbody layouts (from very rough to very clean), ground clearance ( from the 200 mm typical of passenger cars to the 40 - 45 mm of some racing cars).While these tests have certainly provided important information for the specific development of each one of these cars, the possibility of extracting from these results information of more general value is quite limited. In reality, underbody flows are in general quite complex due to the many different details existing on the underbody. Even using the best techniques which are available to map the flow under the car (pressure probes, hot wire, ldv) any attempt to make a complete reconstruction of these flows gives often only modest results. Consequently it is rather difficult to propose significant underbody configuration changes or to establish clear trends of any type, starting from sets of results measured on real cars. 2. On the other hand, as everyone knows, important configuration parameters, like wheel-base, track, front and rear overhangs, wheel-housing size, etc are virtually impossible to be modified during the wind tunnel tests on real cars. Therefore, the potential of aerodynamic improvement related to this type of configuration changes is practically unknown, at least for tests on full size cars and when an appropriate simulation of the ground motion and wheel rotation are considered.The research program started in 1996 at Pininfarina, and partly funded by the Italian National Council for Research, in the “Progetto Finalizzato Trasporti 2” program, intends to investigate this specific field of aerodynamic effects which are associated with important geometric changes of a car underbody. And this, by carrying out tests on a full scale simplified model and using the Ground Effect Simulation System mentioned before, to provide a correct simulation of the underbody flows. A second important aim of the research program is to define, in a systematic way, the differences which can be expected for the various GESS testing conditions.The details of this research program and the results obtained are reported in the following paragraphs.