AERODYNAMICS OF COOLING FOR RX8 TURBO CONVERSION
INTRODUCTION
This is a short article written during the conversion of a Mazda RX8 with twice the power using a 13B REW engine from a Mazda RX7 FD and a single turbo (the FD has twin sequential small turbos). This conversion introduces an extra cooling component, an intercooler, to a cooling system that has to cope with twice the power in the same space.
An engine generates large amounts of heat and an RX8 gets rid of it by 2 oil coolers exhausting into the front wheel wells and a radiator exhausting predominantly over the engine. Some of the radiator exhaust evacuates through the wheel wells and side vents, the latter being of little practical value as it has a narrow and convoluted path (ie the side vents are purely a styling feature. This arrangement is pretty typical of modern cars, combining as it does simplicity, cheapness and ease of manufacture.
Improving this system within the limits of the bodywork to cater for an engine generating over twice the power with a turbo system piling on even more heat is, to say the least, an interesting task requiring a lot of compromises.
Don’t assume that manufacturers have got things right. From cars through to early models of iconic aircraft such as the German JU87 Stuka dive bomber and Bf 109 fighter, designers have things horribly wrong.
We need to cool not only the engine but also air conditioning where fitted, the turbo, gearbox and other items such as the brake fluid reservoir and ignition system that are affected by heat from the engine.
SOME AERODYNAMIC THEORY AND PRACTICE ON DUCTING
Introduction
Most theory and research relates to aircraft engines and there is little beyond the 1940s, when efforts turned to jet engines. Despite the differences of speed and aerodynamic pressures, both theory and practice hold up reasonably well when translating from aircraft to cars, albeit that optimum values change (Reynolds Number Effects).
General Arrangement
Put a radiator in an air flow and the vast majority of the air will spill over the sides of the radiator creating large amounts of drag. Ducting is needed in front and behind the radiator to take only the air required and introduce it back into the air flow. Air will only flow through a radiator if there is a pressure difference across it and efficient cooling needs the air going through the radiator to be turbulent. The least drag and best cooling occurs where the air inlet’s height is 1/6-1/3 of the radiator’s at one radiator height ahead of the radiator and its exit is 115% of this to allow for heat expansion, again at one radiator height from the radiator. The rate of flow is entirely controlled by the exit height. A properly designed aircraft duct will actually generate thrust rather than drag from the Meredith Effect where the radiator heat gives the cooling air extra energy, as was the case for the legendary P51 Mustang fighter.
The ideal way to exhaust cooling air is into the freestream flow in the same direction and speed. In a car this means through a ducted vent in the front of the bonnet where pressures are low and through vents behind the rear wheels. Other places are the wheel wells, over the engine then underneath the body. An often-quoted study of cooling arrangements gives the following analysis:
(i) Speed through Radiator As Propn of Car Speed (ii) Contribution to Cd (Drag Coefficient)
A (Front mount venting into engine bay): (i) 27% (ii) 0.025
B (Front mount ducted to wheel well): (i) 14% (52% of A) (ii) 0.020 (80% of A)
C (Ducted to bonnet/hood): (i) 25% (93% of A) (ii) 0.010 (40% of A)
D (Top-mount inlet ducted to underside): (i) 10% (37% of A) (ii) 0.020 (80% of A)
These figures are indicative only. For example, the way that arrangement C directs the exhaust air has an appreciable effect on the Cd value.
Inlet Configuration
The radiator requires such a large air mass flow that a forward-facing duct is needed, as low down as possible on the front of the car where pressures are highest. Fancy arrangements such as NACA ducts have inadequate air flows and pressure recovery. Bonnet/hood scoops are inefficient and are generally difficult to integrate with possible radiator locations.
The inlet configuration must have a rounded lip and smooth transition along its length in order to avoid flow separation that will choke the duct and hence reduce pressure recovery and increase drag. The simplest duct shape, one with straight sides angled outwards, is known as a rectangular duct; it should have sides angled at no more than 11 degrees to the inlet air flow, which requires long duct lengths of over twice the radiator height.
As outlined earlier, inlet heights of 1/6 to 1/3 the radiator height one radiator height ahead of the radiator give optimum balances of air flow and cooling drag. Surprising to most people, theoretical and practical analyses show that the cooling drag is the same regardless of radiator height, provided the inlet walls avoid flow separation. Larger inlets produce more drag overall because the car’s fixed-size radiator effectively limits the air passing through and so they slow and spill large quantities of air that create more drag than if deflected by a smaller inlet.
The optimum velocity for the air presented to the radiator is a balance between cooling efficiency, where (all other things equal) higher velocities give better cooling, and pressure drop across the radiator (and hence drag) where higher velocities give higher drops and hence higher drag.
Radiator Position
In general, the radiator should be mounted as low down as possible in order to lower the car’s centre of gravity and may be angled by up to 30 degrees without appreciably reducing cooling efficiency. As far as is practical with other considerations, the radiator should be as near to the engine as practical in order to reduce weight, cost and coolant volume.
Exit Configuration
The exit ducting converges to accelerate the exhaust air and therefore has a negative pressure gradient. This means that the rate of change of the duct area is much less important than for the inlet diffuser.
Ideally the exhaust air should be exhausted from an exit perpendicular to the freestream flow. However, we cannot do this. The typical production car method of exhausting through the engine bay underside is relatively inefficient as it creates drag from the exhaust air’s bouncing around off the car’s rough underside. Air could be ducted out of side vents or into the front wheel wells but, in general, the only reasonable alternative to the bottom of the engine bay is the bonnet/hood. The pressures across the bonnet change from very low at the bonnet lip to slightly positive at the windscreen base, so the earlier the exhaust air exits from the bonnet the better.
The diagram below shows 4 alternative designs of exit through a bonnet/hood showing differences in the measures of pressure and drag (Cp and Cd respectively).
In A, B and C the exhaust air flowing along the plate generates lift from its higher velocity (and hence lower static pressure), differential pressure on the shape creates drag and the freestream flow creates a turbulent boundary layer at the junction of the 2 flows, generating drag. In case D, however, the 2 flows mix violently, creating a thick turbulent boundary layer that raises the pressure on the plate after the lower pressure of the exit, creating a roughly zero change in Cp; the additional drag from the turbulent boundary layer is largely offset by the saving in drag from not having differential pressures across a cowl, resulting in only a small increase in drag.
A Gurney flap (basically a short vertical plate) ahead of a bonnet vent will help pull air out of a vent by lowering the pressure behind it at the expense of some extra drag. This is why you see a hump ahead at the front of vents in many carbon fibre bonnets/hoods.
If air is passed across the engine or attached systems such as a turbocharger, it should be exhausted through vents in, by order of preference, the bonnet, bodywork sides behind the front wheel, wheel wells and engine bay underside. Air ducted out of the bottom of the engine bay should be as far back as possible and if practical in line with the car underside.
AERODYNAMICS OF COMPROMISE
The above is all well and good, but the confines of the RX8 engine bay mean that there is inadequate space to duct fully the 4 radiator systems (intercooler, engine coolant, oil cooling and air conditioning). We can rule the air conditioning radiator out as we can either put it immediately in front of the coolant radiator or by removing the air conditioning altogether, so that leaves us with 3 radiators to duct. One possible compromise is to put the oil coolers in a duct just over one cooler height in length venting into the front wheel wells, the intercooler in a duct around 1.6 times the intercooler height and the coolant radiator at the end of and inlet duct 1 radiator height in length and venting over the engine, turbo and gearbox.
I can find no scholarly or practical articles on the best arrangement should the ducting available be less than 2 radiator heights in length. On first principles, I think that we have 2 least-bad alternatives:
• For the intercooler, build the exhaust duct 1 radiator height long and exiting at ¼*115% radiator height and the inlet duct as long as possible with an inlet height matching a ¼ height full length arrangement.
• For the oil cooler, build the inlet duct at ½ length and exhaust through an exit duct shaped to exit air downwards in front of each wheel.
Designing systems for turbo-charged engine in Mazda RX8 and aerodynamic modding for mpg on road and performance on track
Today
Other popular topics in this forum...
