Wednesday 7 November 2012

Radiator theory

I thought I would try to pass on some of what I have learnt and how it relates to what I'm doing with the Twister.

Firstly some basics.

Piston engines generate a lot of heat - approx 70% of the energy generated by a piston engine is heat. Only 30% of the energy goes to actually driving propulsion.

Some heat goes out the exhaust, some is radiated away from the engine itself - the rest must be dealt with or the engine will overheat.

So you can have an air cooled engine or a water (radiator) cooled engine.

An air cooled aero engine is actually better described as an air/oil cooled engine as the oil is a key factor in it's cooling.

If you think about it, all piston engines are air cooled, as it is the air passing over a radiator that is taking the heat away (not the fluid in the radiator itself - that is just moving the heat around).

So to keep the oil cool we need a radiator (as oil is a fluid and this is the best way to cool it).

A radiator without ducting is very inefficient and very draggy.

Only about a third of the air striking its surface actually goes through the radiator. Plus it's whole face is being projected into the airstream resulting in massive drag.

So we must have ducting to be efficient.

Ducting also allows us to put the radiator and it's plumbing out of the direct airstream and expose only it's core to the cooling air.

So where to put this radiator?

Pete has put his in front of the engine as you can see below.

He has been through about 4 variations to get to this point where his temps are okay. (He's also had to put some more exit louvres on the rear of the cowling).




















So although this is a known working solution there are two things I don't like about it.

1. It's inefficient.

2. It's ugly.

Note the size of the inlet is the same size at the radiator face - therefore he is not getting any of the efficiencies that a duct brings. There is no outlet duct on his setup.

By having a trumpet shaped inlet the air inlet opening can be much smaller than the radiator face - reducing drag.

A duct shaped like a trumpet at the radiator face does two things.

1. It slows the air down - thereby allowing it more time in contact with the radiator as it passes through it - so better cooling.

2. According to Bernoulli if the velocity is reduced the pressure must increase (the opposite of what happens on the top of a wing). Therefore if the pressure goes up then this means more air molecules passing through the radiator and so better cooling. It also means that the air is more inclined to go through the radiator (to an area of lower pressure.)

See this diagram below for the ideal inlet and outlet shapes - note that the lower funnel shape is incorrect. And the outlet at the top has a nozzle at it's end. (more about that later).



















As Pete has a retractable gear version there is nowhere else to put the oil cooler so I can understand why he has put it where it is.

As I have the fixed gear version I have a big empty space in the fuselage that would normally be taken up by the main gear.

Not only that, this location is further back on the airframe - anything further back on the airframe suffers less drag as the air has already been disturbed before it has got there. (A generalisation).

Also under the wing there is an area of high pressure - just what we need for an inlet.




















Not only that - this high pressure moves rearward when the angle of attack is increased - such as when in a climb attitude. (see above diagram).

This puts the air where we need it when we need it most - in the climb.

For the air inlet scoop I was influenced by the P-51 Mustang.

















Notice the position on the airframe - under the wing and down away from the boundary layer. Plus the opening is angled rearwards so that maximum air enters in a high angle of attack - such as in the climb.

For the exit I have positioned the outlet as far back as I can where the wing has an upward slope. I am hoping that this provides a suction type effect as well as a very smooth exit of the airflow back to the surrounding airstream (this is the ideal situation).

Just recently I found this diagram of the De Havilland Mosquito radiator. The radiators are positioned in the leading edges of the wings.

Note the exit - which is what I am doing - although I'm sure the designers would have liked to exit it further back but the main spar of the wing prevented this.

Note also the cowl flap which does two things.

1. Controls the exit volume.

2. When open it makes a 'lip' that generates low pressure behind it - effectively sucking the air out.

I will also have a cowl flap and I expect to have to use it in the open position on the ground and whilst in the climb or doing aerobatics. When on a reduced power setting at a higher speed in the cruise I hope to be able to close it completely.


















Now I know you're thinking - how about some modern designs that have moved on from WWII? Well, believe it or not things never really progressed after WWII as the jet age took over and all the best designers worked on jets after that so very little research and development in this area was done post WWII.

The exit design is the crucial bit to recover the energy losses from the drag caused by the inlet and the air's passage though the radiator itself. Most people don't realise how important the exit is.

In 1935 an Englishman called F. W. Meredith (working at RAE Farnborough) wrote a paper on this and ever since it has been known as the Meredith Effect.

Basically the idea is to heat up, then speed up the airflow as much as possible to, in effect, create a little thrust (a bit like a mini jet engine).

I don't expect to get much (if any) Meredith effect from my set up as the Twister does not fly fast enough. Generally anything over 300mph and the Meredith effect becomes a big deal. The Spitfire was the first aircraft to utilise the effect and it counted in a big way towards it's top speed.

However the Spitfire designers/developers never made the most of the effect as the ducts were too short. The Mustang made full effect of it and reduced it's cooling drag from 400lb to 50lb (negative thrust) - giving it a great top speed and helping it have a tremendous range.

Note the net outcome of the Meredith Effect is not thrust but drag reduction.

The principle remains the same though - speed up and direct the exit air to as close to the free airstream as possible to minimise losses.

Below is my set up in the Twister.






























As you can see the theory is all there - however as always with these things the proof will be in the test flying - we will see...

4 comments:

  1. Richard Caldwell has left a new comment on your post "Radiator theory":

    Hi. Nice info, especially the diagram of the Mosquito's system. I like your solution and would like to suggest an improvement. The back of the intake fairing should be extended to merge with the exhaust, possibly with vertical flow-separators extending to the back of the exhaust to keep side airflow from interfering with things.

    ReplyDelete
  2. Hi Richard, thanks for your comment. If the inlet shroud were to extend all the way back to the exhaust outlet then this would increase the wetted area dramatically, and thereby increase the drag of the shroud by an unacceptable amount. Although it is not shown in this post, the exhaust exit flap has side vanes which prevent air from entering or disturbing the exhaust air when the flap is open. See my other posts on fabricating these parts to see what I mean. What is your experience with ducts? And which aircraft are you building?

    ReplyDelete
  3. Hi, your 'diagram below' needs the tick and x swapped. The trumpet diffuser will seperate at the steep part, effectively blocking the outer edges of the cooler. Best is a straight sided diffuser of max 7 degrees. That can be a pain to fit, but you can divide it up into multiple sections with dividers. Don't try turning the air sharply while trying to diffuse it. Separation again. Many compact ducted coolers are inclined in the duct. Can work well. Lots of small turning vanes can make this pretty efficient. Compared to the inlet, you can get away with murder in the nozzle. A straight sided cone has the lowest losses, though. A 30 degree angle is about optimum.

    ReplyDelete