Reduction of Aerodynamically Undesirable Influences Due to Engine Cooling Flow in Road Vehicle

The purpose of this research is to clarify causes for the change in aerodynamic characteristics of a road vehicle model due to engine cooling flow in wind-tunnel experiments with the moving-belt ground board, in order to propose methods to reduce the drag and lift. With regard to engine cooling flow, the air-intake system was adjusted with variable opening area and position for the engine loading system of FF and FR with and without a radiator. A simplified 1/5 scale vehicle model was manufactured with transparent externals around the engine for flow visualization. The overall results show that with enlargement of the opening area, the drag and the front lift increased and the rear lift decreased. The flow visualization and the measurements of underfloor velocity and surface pressure indicated the cause of the characteristics changes. Enlargement of the opening area causes flow disturbance by merging of the scavenging flow and the underfloor flow, which has blockage effects for the upstream of each flow with keeping high pressure in the engine compartment and causes pressure loss under the floor behind the engine unit. The difference between the two engine loading systems lies indirection and location of the engine unit, which causes the differences of how the flow features affect the aerodynamic characteristics. The effect of the radiator is to reduce the range of changes in drag and lift. Finally, it is discussed that the principle of reducing drag and lift is to suppress interference of scavenging flow, and concrete methods are proposed.


Introduction
tant topics due to aggravation of the global warming and the energy problem, and therefore reduction of the aerodynamic drag that greatly influences the fuel efficiency is proceeded [1]. Aerodynamic drag basically consists of two types, i.e., pressure and friction drags. As for road vehicles, more practical classification is made such as a shape drag (mainly pressure drag), an internal-flow drag, an induced drag, an interference drag and a friction drag. Since the shape drag occupies 75% of aerodynamic drag of whole car, optimization for configuration has been regarded as most important in aerodynamic developments of cars [1] [2] [3]. History of improvements of the car shape dates back to the 1920s'. The average for the coefficient of drag at that time is about 0.8, but nowadays that for passenger cars is about 0.3, which shows the evolution of car shapes during about 100 years. Recently, in order to realize further reduction of aerodynamic drag, attention has focused on reduction of not only the shape drag but also the internal-flow drag. It is generated mainly when the wind blows through the front air intake to the engine compartment for cooling a radiator, an air-conditioning condenser and an engine itself. The internal-flow drag through an engine compartment occupies about 10% of the whole aerodynamic drag, and today when the shape of the car body has improved, the reduction of the cooling-flow drag begins to be put into emphasis.
To consider the cooling drag, Wiedemann assumed the engine compartment as a huge duct system theoretically [4], Barnard, et al., found experimentally that the outlet angle and location of the duct are critical [5], and Braeder, et al., used a generic car model with a simplified internal duct and showed numerically and experimentally that the cooling air interferes with the external flows, designing a model radiator for pressure loss of the radiator, engine block, etc. [6]. On the other hand, Nouzawa, et al., considered the engine cooling air in the system of a real vehicle consisting of: 1) flow around the front-end shape; 2) flow between the grille and the radiator; and 3) flow within the engine compartment and scavenging, with experimental analysis [7]. Recently numerical simulation on the cooling drag has been carried out in the system as each real vehicle [8] [9] [10]. In addition, with regard to the engine-compartment aerodynamics, other than documentations about drag and cooling efficiency, there are few publications about the running stability such as reduction of lift and about the side-wind stability, and therefore further research will be necessary. The authors investigated effects of the aerodynamic drag and lift by the engine cooling flow in a simplified vehicle model experimentally with variation of the area and position of an air intake, the engine layouts for FF (Front engine & Front drive) and FR (Front engine & Rear drive), and existence and non-existence of a radiator, and concluded that the drag and lift are influenced overall by the intake area [11].
In this paper, for the engine loading systems of FF and FR, both of which are typical drive systems in mass-produced vehicles in Japan, causes for the above change in aerodynamic characteristics are investigated by smoke visualization and measurements of the flow velocity and its RMS (Root Mean Square) under

Test Model
The design intent of the test model in this study was to simulate the engine cooling flow that can be applied to almost all vehicle types with a front engine by simplifying common items.
A simplified 1/5 scale vehicle model with specifications of Table 1 was produced [11] based on the average dimensions of typical domestic vehicles in Japan. The whole view and detailed dimensions are shown in Figure 1(a) and : one is an engine mounting method called "width placement" for FF cars, and the other is that called "length placement" for FR cars. In the former the unit is located almost at centre of the engine compartment, while in the latter the transmission part is attached to the back board of engine compartment. The engine unit is clamped to the lower main frame.
The front plate is replaceable so that the air-intake opening can have six sorts of inlet heights from 0 mm to 100 mm shown in Table 2  Further, by measuring the loss of pressure coefficient (C p ) in a real vehicle radiator, a radiator model with equivalent C p loss was fabricated by stacking eight steel nets (Figure 4(c)), and installed immediately after the front main frame as       shown in Figure 3. Cases without and with a radiator (two ways) were considered.
In the width-placement engine (Figure 3(a)), the airflow coming from the intake opening escapes to the underfloor of the vehicle body through the front, rear and sides of the engine unit, or out of the fender through the wheel housing.
In the length-placement engine (Figure 3(b)), there is no spillage through the rear of the engine.

Experimental Equipment and Measurement Methods
Experiments were carried out by using the large-scale low-speed wind tunnel of  [12] and setting the moving belt ground board with a boundary layer suction device between the nozzle outlet and the collector of the wind tunnel.
As indicated in Figure 5(a) and Figure 5(  semiconductor pressure transducer MS4515 (Measurement Specialities Inc.) was adopted, with sampling period of 2 ms and sampling number of 1024. This device was produced to measure the surface pressure on car models, and the validation was shown in [13] [14] where the unsteady characteristics of circular-cylinder flow are captured.
In each measurement of force and pressure, an average value for sampling number 1024 was calculated, which was repeated three times, and further the average for three times is shown with error bars of maximum and minimum averages as the results.

Experimental Conditions
The test model covers 48 combinations from the six heights with the lower or upper position for the air intake, existence and non-existence of the radiator, the two types of engine loading systems. The road clearance for the test model is 30 mm.
The flow conditions were 20 m/s for both the wind speed and the moving-belt speed. It has been confirmed that uniform flow is maintained up to the vicinity of the ground board by the moving belt with boundary layer suction at measurement points shown in Figure 6, not only along the belt center but also up to the vicinity of the belt side ends [15]. Typical velocity profiles are shown in

Results
In the force measurements in Section 3.1, results for both the lower and upper positions for air-inlet opening are shown, whereas in Section 3.2 and later, results for the lower position only are shown, as it is taken in much in domestic passenger cars.  In Figure 7(a), with enlargement of the inlet height, C D increases overall, and in Figure 7(b) by equipping the vehicle model with the radiator, the increase tendency is relaxed, and the increase is slighter in the engine loading system of length placement. This leads to the consideration that in the width-placement engine, the inlet opening area has a dominant effect for the drag, while in the length-placement engine, existence or non-existence of a radiator is dominant.

Aerodynamic Characteristics
In Figure 8(a), with enlargement of the inlet height, C LF increases. Further in Figure 8(b) by installing the radiator, the difference due to setting conditions of engine loading system and opening position becomes smaller, and the increase tendency is relaxed with rise at 0 mm and 20 mm in inlet height and fall at 80 mm and 100 mm. By full opening, increase of about 0.5 without the radiator and about 0.2 even with the radiator is observed.
In Figure 9(a), with enlargement of the inlet height, C LR decreases, with difference due to the engine loading system and the opening position. In Figure  9(b) with the radiator, the decrease tendency and difference due to the setting conditions is relaxed.      In Figure 10 for the width-placement engine, in (a) for opening height 20 mm, fluid entering from the inlet into the engine compartment flows up to the engine top surface, while fluid entering from the underfloor behind the engine unit winds up to the engine top. The two-course flow meets on the engine top surface and slightly exits through the wheel housing. In (b) for opening height 80 mm, fluid entering from the inlet runs up, passes over the engine top, and flows out behind the engine unit to the underfloor, while fluid entering from the lower side of the inlet flows out of the engine front to the underfloor. A slight outflow through the wheel housing is also observed. Compared with (a), winding up behind the engine unit is not identified.

Schematic Flow Patterns
In Figure 11 for the length-placement engine, comparing the cases of opening height (a) 20 mm and (b) 80 mm, fluid entering from the inlet into the engine compartment flows up and passes over the engine top with disturbance, regardless of the inlet height. Regarding the outflow, in (a) for height 20 mm, outflow only through the wheel housing to the outside of the fender was identified, whereas in (b) for height 80 mm, fluid passing over the engine top flows out through both sides of the transmission to the underfloor and also through the wheel housing to the outside of the fender. Fluid entering from the lower side of the inlet flows out of the engine front directly to the underfloor.
Cases (c) in Figure 10 and Figure 11 show oblique top view for opening height 80 mm when the smoke is generated at a side corner of the front end. Comparing the smoke top views, it is observed that smoke flows to outer side of the fender more in the length placement engine, which suggests that the outflow rate from the wheel housing is higher in the length placement engine than in the width placement engine.   increases and that by equipping the vehicle with the radiator RMS distribution is relaxed. That is, in inlet height 20 mm, the local minimum of velocity at 29 cm distance from the front end and the local maximum of RMS at 24 cm disappear by installation of the radiator, while in inlet height 80 mm, by the radiator, the velocity distribution does not change clearly but the RMS is lowered after 24 cm distance. Figure 14 for the length-placement engine shows that tendency of both velocity and RMS distributions is divided by the presence or absence of the radiator.    In the cases without the radiator, as the inlet height increases, (a) the flow velocity is lower and (b) its RMS is higher under the engine compartment. In the cases with the radiator, for increase of the inlet height, the peak velocity at 4 cm Journal of Flow Control, Measurement & Visualization slightly decreases, and the maximum RMS does not change but transference of peak position from 24 cm to 34 cm distance makes RMS larger in range from 29 cm to 60 cm in inlet height 80 mm. By comparing Figure 13 and Figure 14, it is observed that in the lengthplacement engine the maximum velocity is higher and the peak RMS is mostly lower than those in the width-placement engine. Next, we compare results in more detail between width-and length-placement engines in Figure 16 and Figure  In addition, at the rear end of the vehicle body (point Nos.27-31) it was confirmed by measurements that the difference of surface pressure due to change of the inlet height and the radiator is small.

Pressure Distribution about Engine Compartment
About the engine-transmission unit, pressure holes are arranged at intervals of Journal of Flow Control, Measurement & Visualization     In Figure 18 about the width-placement engine, overall tendencies can be In Figure 19 about the length-placement engine, overall tendencies are as fol- Comparing the pressure values in (a) to (c) between Figure 18 and Figure 19, the difference appears prominently in inlet height 80 mm: in the length placement the pressure is higher with larger error ranges in absence of the radiator and is lower in presence of the radiator than in the width placement

Discussion
We discuss effects on enlargement of the air-inlet height by comparison between inlet heights 20 mm and 80 mm.

Engine Loading System: Width Placement
In the width placement as engine loading system, same tendencies for the Journal of Flow Control, Measurement & Visualization Figure 19. Pressure distribution about engine compartment with length-placement engine.
change of the air-inlet height have been shown in each case with and without the radiator. First, we consider causes for increase of C D due to enlargement of the inlet height shown in Figure 7. As the inlet height is enlarged, the pressure within the engine compartment increases (Figure 18), and the drag increase is due to increase of the pressure difference between the front and back of the engine unit and also due to increase of the pressure on the back board behind the engine unit. Further, the pressure on the engine bottom increases and so does the pressure at the frontward side, which is considered to be connected with the pressure on the front end of the vehicle body, and accordingly the pressure on the front window also increases (Figure 16(b)); thus the drag increase is also considered to be due to high pressure on the front end and front window of the vehicle body induced by high pressure inside the engine compartment. The pressure on the rear window and on the rear end does not largely affect the drag, because the pressure change was very small.
Regarding flow about the engine compartment, with enlargement of the inlet height, the outflow from the engine compartment toward the underfloor increases ( Figure 10), which causes disturbance of flow (Figure 13 Next, we consider causes for increase of C LF with enlargement of the inlet height shown in Figure 8. It is due to the pressure rise under the engine unit ( Figure 18(c)), the reason of which has been discussed in the cause of the increase in drag. Also, as shown in Figure 16 Finally, decrease of C LR with enlargement of the inlet height ( Figure 9) is considered to be due to the pressure decrease under the floor of the vehicle body behind the engine compartment (Figure 16(a)). As discussed in the cause of the increase in drag, the scavenging flow causes the flow disturbance, resulting in energy loss behind the engine compartment, which leads to the pressure fall. The pressure on the rearward of roof and on the rear window does not largely contribute to the rear lift because the pressure change due to enlargement of the inlet height is very small. Furthermore, the reason why the radiator reduces the range of changes in drag and lift in Figures 7-9 is discussed below. In each of inlet height 20 mm and 80 mm, the pressure distributions in Figure 16 and Figure 18 change so that the opening height decreases due to the attachment of the radiator. This suggests that the passage inflow rate is reduced by installing the radiator, and in fact, a slight flow escaping downward from the front of the radiator and slower flow velocity after the radiator was observed by the smoke method (see Section 3.2). The drag change in Figure 7 shows the effect of reducing the opening height by the radiator. However, the changes of front and rear lift in Figure 8 and Figure 9 show the same reduction effect for the larger inlet height, but show the opposite effect for the smaller inlet height. The reason of the latter is considered that by installation of the radiator, the downward momentum of the scavenging air from the engine compartment increases, so that the front lift increases, and the rear lift decreases due to the moment balance around the support point. Journal of Flow Control, Measurement & Visualization

Engine Loading System: Length Placement
In the length placement as engine loading system, first, we consider causes for increase of C D due to enlargement of the air-inlet height shown in Figure 7.
Without the radiator the drag increases as in the width placement engine, but with the radiator the increase is slighter. Also from distributions of the underfloor velocity and RMS ( Figure 14) and the engine-unit surface pressure ( Figure   19), it is considered that the presence or absence of the radiator has a great influence on the drag, above all, at inlet height 80 mm in the length placement.
As the inlet height is enlarged, the pressure on the front and top surfaces of the engine unit increases uniformly in space (Figure 19). Without the radiator, at inlet height 80 mm, pressure is significantly high; above all, the pressure on the engine top is higher by 100 Pa than the highest value in the corresponding case for the width placement engine. The reason is considered that the high stagnation pressure on the rectangular engine front is maintained near the engine unit because the upper space of the transmission is almost blocked by the engine, the bonnet and the back board. Even with the radiator, the pressure also increases due to the inlet height although the rise is very small. Thus, a cause of the drag increase is the pressure rise on the engine front and on the back board.
About the pressure around the vehicle body, first the pressure under the engine unit (Figure 19(c)) is examined. As the inlet height is enlarged, without the radiator the pressure on the engine bottom increases, whereas with the radiator it decreases and in inlet height 80 mm the pressure shows the lowest with the pressure gradient smaller. It is considered that on the front end of the vehicle body the pressure does not vary much due to existence or non-existence of the radiator. The pressure on the front end would affect the pressure on the front window, which accordingly increases without and with the radiator (Figure   17 Figure 11). In absence of the radiator the underfloor velocity decreases with increase of its RMS (Figure 14), and the pressure under the engine compartment increases (Figure 19(c)). This is considered to come from the effect of blocking the underfloor flow by underfloor disturbance due to the engine scavenging flow, as discussed in Section 4.1. On the other side, with the radiator, the reason for the decrease of pressure on the engine bottom (Figure 19(c)) is expected that under the floor the RMS decreases and the underfloor velocity increases. In Figure 14, however, such tendency for RMS and velocity does not clearly appear. This might be because the velocity measurement was carried out for horizontal direction only. In addition, due to the positional relationship be-T. Sawaguchi, Y. Takakura Journal of Flow Control, Measurement & Visualization tween the wheels and the engine, the high pressure in the vicinity of the engine unit allows airflow to easily escape outward from the wheel housing ( Figure 11).
Summarizing the above, increase of the inlet height causes significantly high pressure on the engine unit due to the forward block by the length-placement engine, and the scavenging flow merges with the underfloor flow to make disturbance, which has blockage effects for both the scavenging flow and the under-

Comparison between Width and Length Placements of Engine
Overall The differences of phenomena between the two engine loading systems arise from placement direction (width or length direction) and location (distance from the air inlet to the engine front) of the engine-transmission unit inside the engine compartment. Here, focusing on inlet height 80 mm where remarkable difference appeared, we discuss the comparison.
In the length placement without the radiator, fast flow runs against the square shape of the engine front with higher stagnation pressure generated than against the notched square shape of the engine-unit front in the width placement ((b) in Figure 19 and Figure 18), and due to the shape almost surrounded by the engine unit and the backboard, the pressure becomes much higher in top of the engine unit ((a) in Figure 19 and Figure 18). However, the pressure on the front end of the vehicle body is considered to be lower, since the distance from the air inlet to the engine is longer and through the wider space the fluid easily flows out more from the upstream side of the engine to the underfloor and to the outside of the fender with induction by the rotation of wheels ((b) in Figure 11 and Figure 10). The lower pressure at the front end would induce the lower pressure distribution on the front window ((b) in Figure 17 and Figure 16). Thus, the pressure is higher in the vicinity of the engine unit but lower on the front end and the front window, which leads to the drag slightly smaller in Figure 7(a) without the radiator, at inlet height 80 mm with lower position, compared with the width placement. By installing the radiator in the length placement, the pressure on the engine front is considerably lower than in the width placement ((b) in Figure 19 and Figure 18). The reason is considered that, in the length placement the slow stream behind the radiator flows out more from the upstream side of the engine to the underfloor and to the outside of the fender as stated in the previous paragraph, so the velocity of flow in the straight direction would be further lowered, and the stagnation pressure on the engine front becomes lower. The pressure on the front end of the vehicle body would not change much due to the radiator as stated in Section 4.2, and so the pressure on the front window would not. Thus, the lower pressure in the vicinity of the engine unit and on the front end and window leads to the drag quite lower in Figure 7(b) with the radiator, at inlet height 80 mm with lower position, compared with the width placement.
As the factor of the drag increase due to enlargement of the inlet height, in the length placement, the change of pressure in the vicinity of the engine unit is stronger without the radiator and the change of pressure on the front end and window of the vehicle is weaker than in the width placement.
In the width placement, since there are spaces in front of the engine unit and behind that, the scavenging flow can run out of the spaces to the underfloor, while in the length placement, since there is no space between the transmission and the backboard, the scavenging flow runs out of the front side of the engine and both the sides of the engine unit to the underfloor, and runs away out of the wheel housing to outside of the fender ((b) in Figure 10 and Figure 11). Therefore, in the width placement, the outflow rate of scavenging to the underfloor Journal of Flow Control, Measurement & Visualization would be higher and the disturbance is generated more overall under the floor ((b) in Figure 13 and Figure 14), with energy loss higher, resulting in lower underfloor pressure in the downstream of the engine compartment. In comparison of (a) between Figure 16 and Figure 17, in the vicinity of hole No.1 the pressure is remarkably lowered on the underfloor center line, although the point would contribute to the front lift. Not only on the center line but also on the whole underfloor surface the pressure loss is considered to spread more in the width placement. The lower underfloor pressure leads to the lowest rear lift at inlet height 80 mm with lower position in Figure 9(a) and Figure 9(b), without and with the radiator.

Reduction of Aerodynamically Undesirable Influences
From the previous Sections 4.1 to 4.3, the principle to simultaneously decrease drag and lift is considered to avoid interference of the engine-compartment scavenging flow with the underfloor flow coming from the upstream side to reduce disturbance under the floor, which suppresses pressure rise inside the engine compartment, and accordingly on the front end and front window of the vehicle body. In underfloor scavenging, scavenging toward the rear of the vehicle body is considered effective to avoid interference between the engine cooling airflow and the underfloor. One example is attachment of an engine undercover [8].
Other than underfloor scavenging, it would be also effective to perform upward scavenging from the bonnet or sideward scavenging from the fender.
In the length placement without the radiator, since the pressure in the vicinity of the engine unit is remarkably high, above all, upward scavenging from the bonnet is considered effective by making the ventilation port in the bonnet, because the pressure inside the engine compartment is lowered without underfloor disturbance. One example has been demonstrated by the authors [16].

Conclusions
In this experimental study, the test model was designed to simulate the engine cooling flow that can be applied to almost all domestic vehicle types with a front engine by simplifying common items, and for change of intake opening area the changes of aerodynamic characteristics were investigated with variation of the intake position and the two engine layout of width and length placements in presence or absence of the radiator. The overall results show that with enlargement of the opening area, the drag and the front lift increased and the rear lift decreased. In this model, flow separates at the front corner of the simplified vehicle body but reattaches on the front window. Therefore, it is considered that the tendency of change in drag and lift is universally applicable to real road vehicles without separation at the rounded front end, but with lower drag.
Here it has been clarified that the cause of the change in aerodynamic characteristics due to the cooling airflow is disturbance under the floor due to merging of the engine scavenging flow and the underfloor flow. In the correlation of Journal of Flow Control, Measurement & Visualization phenomena, enlargement of the air-inlet height (and therefore increase of the cooling flow rate) causes pressure rise inside the engine compartment, and further adapts to that on the front end and window of the vehicle, but causes pressure fall under the floor behind the engine compartment. It is due to the flow disturbance above stated, which has blockage effects for each upstream of scavenging flow and underfloor flow with keeping high pressure in the vicinity of the engine unit (except for the pressure on the bottom of the length-placement engine with the radiator) and causes pressure loss in the downstream of the merging disturbance under the floor.
Further, it has been shown that differences in flow features and aerodynamic characteristics between the two engine loading systems arise from direction and position of the engine unit inside the engine compartment. Since the pressure inside the engine compartment is connected with that on the vehicle surface by the pass way of engine cooling flow, the principle to improve drag and lift is considered to avoid interference between the scavenging flow and the underfloor flow. Other than underfloor scavenging without interference, upward scavenging from the bonnet and sideward scavenging from the fender is considered effective. In the length placement engine without the radiator, since the pressure on the engine unit is remarkably high, upward scavenging is recommended.