Computer simulation of vehicle performance

[Pumaracing]

Pumaracing said:

stevesingo said:

Dave/Stan,

Can you do the calc for the following…

Flywheel bhp is 340. Wheel bhp is 289.

So? What is it and what power do you think you have?

[Pumaracing]

ivanhoew said:

a few different chassis dyno runs ,at different places ..

Phew. Lots of data. Struggling through it all though it doesn't appear to be a million miles out from the simulation prediction of 213 bhp at 19psi.

Losses in Mini engines are actually quite high with the drop gears but of course you won't necessarily see those on the overrun on a chassis dyno when the engine isn't producing power. I've factored those in when I said 180 ish wheel bhp.

[Pumaracing]

PeterBurgess said:

Hi Stevie

The guys who have done best in terms of championships with our heads/engines/dynoing since 1987 have prepared cascade graphs based on power curves in each gear. This can and does give rise tochanging gears at different rpms to hit best torque in each gear.

I attach a pic showing bhp/torque against mph. This can be converted to rpms in each gear. Quite often don't need to rev so high to change into higher gears.



Peter

I'm not sure what you think you're graphing there but it's complete bks. The power and torque should be the same in each gear obviously. The engine doesn't change when the gears change.

[Pumaracing]

stevesingo said:

Apologies for not responding sooner.

The data was derived from GPS, so not that great really, and the run was done on the road so the traction is not so good and it it quite tricky to get off the line.

It makes 237 RWHP 280 Fly on a DynoJet dyno.

Hmmmpf. Now he tells us. That's why the data looked so bad and the speeds were too fast for the times. If you really have about 280 bhp then you'd struggle to hit 104 mph in the 1/4 on a proper flat track.

[Pumaracing]

stevieturbo said:

But often a lazier run can still yield good trap speeds, sometimes higher over that same fixed distance. But it will take longer. ie more time on track to achieve a speed.

This doesn't make any sense at all from a physics perspective. You can only spend more time on the track by going slower on average.

However, the first few feet of track are not that important in terms of the terminal velocity at the 1/4 mile. Think about it this way. If you mess up the launch or deliberately take it easy what you've effectively done is lost a few yards at the "end" of the track in which to accelerate. But acceleration rate at the end of the track is very low so not much speed is being added there so it is of little detriment to terminal speed.

Think about it a second way. You dump the clutch too hard at the start and just sit there for 2 seconds burning rubber going nowhere. Then you lift off a touch, grip, and do a completely normal run from then on. Your ET will be 2 seconds slower than usual but your terminal speed will be the same as any other proper run. The time you wasted didn't use up any track distance.

So no, you can never reach a higher terminal speed by deliberately going slowly somewhere but you won't hurt trap speed much by being a tad slow off the line. However you can NEVER increase trap speed doing that.

[Pumaracing]

ivanhoew said:

i wonder if you could do me 3 sims dave,

1) 1/4 mile with some yokos and a lsd ...eg decent grip/launch , say a 1.9 60 ft ?

Even if you had the grip you don't have the power to go that fast. You ideally want a 3.4 diff with the 12" tyres because it's not revving high enough in 4th. However if I put bhp up to 230 at 23 psi, 3.4 diff, no grip limit I get 60ft in 2.0, 1/4 in 12.1 @ 114.4.

ivanhoew said:

2)max power top speed with the right gearing , i can run down to a 2.95 diff .

Even that's not really low enough. Depends what your rpm limit is. About 139 mph at 19 psi, 7180 rpm. Maybe 141 mph at 23 psi, 7281 rpm.

ivanhoew said:

3) top speed with present 3.1 gearing and 117 bhp at wheels , my normal rd 5psi setting.

122 mph @ 6620 rpm.

[Pumaracing]

Stan Weiss said:

Steve,
If that run had been done on a proper drag strip where you had some roll out (I used 11.75 inches) I get the following still using my 242 HP.

60 Foot ET = 2.6765
330 Foot ET = 6.3456
1/8 Mile ET = 9.1775
1/8 Mile MPH = 86.9702
1000 Foot ET = 11.6147
1/4 Mile ET = 13.6606
1/4 Mile MPH = 110.8751

Stan

We have some serious discrepancies. I know we have the same basic maths in there for aero and rolling resistance drag, tractive force at the wheel and net acceleration so it must be in the details which are crucial to simulating a real car. So here is all my data for this car and the times calculated.

No rollout
weight with driver 2860 lbs
frontal area 20.24
Cd 0.33
tyre radius 0.9689
diff 3.15
gears 3.72,2.4,1.77,1.26,1.0
Transmission efficiency 0.86 in all gears
Wheel/tyre inertia 125 lbs
Engine inertia 25 lbs at an average radius of gyration of 3.5 inches
Clutch slip rpm 5000
Peak rpm 8000
gearchange time 0.4 seconds
Max grip 0.55g

flywheel power curve
4500 - 155
5000 - 185
5500 - 217
6000 - 235
6500 - 260
7000 - 270
7500 - 280
8000 - 265

60ft - 2.60 @ 31.42
110 yds - 6.33 @ 65.46
220 yds - 9.33 @ 83.72 (82.17 @ 210 yds)
440 yds - 14.03 @ 104.93 (104.22 @ 430 yds)
Top speed 169 @ 7695 rpm in 5th

See how you get on.

[Pumaracing]

stevieturbo said:

200hp isnt enough to get a sub 2.0s 60ft on a Mini ?

I thought you were the one who just said that 2.9s is good for a rwd car on road tyres? How is sub 2.0 easier on a fwd on road tyres?

[Pumaracing]

Stan Weiss said:

Dave,
Using what parameters of yours that I have in my software and making an adjustment to one of them for my software this is what I get.

Much better

Makes rather a difference doesn't it having an appropriate power curve rather than something from a Chevy V8 peaking 2000 rpm too low and a gear change time that a human being can't actually achieve. Now you just go right ahead and give yourself a pat on the head and you can open a bag of Maltesers as a special treat.

Computers are great except of course for the proviso that GIGO!

I'll cover rotating inertia briefly.

1) Wheel/tyre inertia can be considered as part of the base vehicle mass as it is independent of gear ratio. Obviously it has to be kept out of any downforce or braking calculations though. After much buggering about with angular momentum and acceleration calculations I concluded that the effective mass add-on is approximately 75% of the total real wheel/tyre mass. I use 125 lbs as an average for road cars and adjust as necessary when I know the wheels will be exceptionally heavy or light.

2) Engine inertia. The equations in each gear with units in lbs and feet are as follows:

(FD ratio x gear ratio x radius of gyration / tyre radius)^2 x effective mass at the specified radius of gyration.

For general car engines I take the average radius of gyration of crank, flywheel and pistons as being 3.5" (0.29167 ft). I adjust this for engines with very small clutches and flywheel, short crank stroke etc like motorbike and F1 engines. Normally down to 2.5".

For effective mass for road type engines I use 25 lbs for 4 cylinder, 30 for 6 and 35 for 8 cylinders. For F1 engines I use 15 lbs.

Therefore for the BMW in question the calculation in 1st gear is (3.15 x 3.72 x 0.29167 / 0.9689)^2 x 25

= 12.44 x 25 = 311 lbs.

The effect drops off rapidly in the higher gears, 2nd 129 lbs, 3rd 70 lbs, 4th 36 lbs, 5th 22 lbs. However it's vital to getting the launch parameters right for the 60 ft and 330 ft times.

So the total effective accelerated mass for the BMW in 1st gear is 2860 lbs (1300 kg) base mass + 125 lbs wheel/tyre inertia + 311 lbs engine inertia = 3296 lbs

[Pumaracing]

ivanhoew said:

brilliant thank you , i know one of our turbo mini chaps ran a 1.9 60 ft , on a 4.11 diff and yoko 32's ,was a bit lighter that my car i think ,

Pumaracing said:

122 mph @ 6620 rpm.

i have had the car at 6700 on that boost , it was just about topped out , so that fits nicely . thank you dave , very interesting .

regards
robert

Remember that on a drag strip the 60 ft times can be artificially reduced by up to 0.35 seconds by roll out before the timing light is actually triggered. I have no truck with such foolishness and my program calculates the actual real elapsed times from the instant the clutch bites.

So my 2.0 second unlimited grip scenario could be about 1.7s on a strip and it's vanishingly unlikely you'll get anywhere near that without slicks but I used it to give you an idea of the best possible 1/4 mile run you might ever be able to do. Anything within 0.3 or 0.4 seconds of what I calculated would actually be very good going for a real car with the usual homo/mechanical piloting interface

[Pumaracing]

ivanhoew said:

i have had the car at 6700 on that boost , it was just about topped out , so that fits nicely . thank you dave , very interesting .

regards
robert

The actual unlimited distance top speed the computer calculated was 123 mph at 6675 rpm but it took 2 miles to get there. I knocked it back by 1 mph to 122 which was more achievable at about 1.25 miles. Headwinds, tailwinds, slopes can have a massive impact though. My program can allow for those too.

Just a 1 degree down slope would add 2 mph to top speed as would a 5 mph tailwind. With both together it would hit 128 mph at 6946 rpm.

So it's a bit of a lottery trying to precisely determine top speed and hence bhp on the road but we appear at least to be very much in the ballpark.

I built in a bunch of other stuff to the program over the years. It can cope with downforce for F1 type cars which of course drastically increases traction at higher speeds. It does braking time and distance calculations based on not just tyre grip but also factoring in aero drag and finally by combining both it can calculate the effect on lap time of bhp or gearing changes by simulating a notional track comprised of given straight lengths and corner entry speeds. It's been fun playing with it and writing it anyway.

[Pumaracing]

Some of you might find this thread on another forum of interest. I used my simulation program to help calculate the power of 2014 F1 cars from the onboard speed display.

http://www.f1technical.net/forum/viewtopic.php?f=4...

We concluded based on not very ideal data that it was somewhere in the region of 700 bhp at the wheels including the 160 bhp hybrid power unit. Maybe 790 flywheel.

[Pumaracing]

Stan,

I think I've found the bug in your program. It's in the rollout section. I input the torque curve you first tried ending at 6750 rpm 168.64 ft lbs, your tyre rolling radius, your gearchange time of 0.1 seconds, took out all the inertia data and still only got a terminal speed of 106 mph at the 1/4. You got 110 mph.

By my calculations a 12" rollout might reduce the 60 ft ET by about 0.3 to 0.35 seconds and the speed at 12" is about 4 mph at 0.5g.

HOWEVER! that 4 mph does not add on to every trap speed up the run. It just means the 1/4 mile is now 1321 ft instead of 1320 ft and the gain in speed is immaterial. The car is hardly accelerating by the end of the run and one extra foot is just a tiny fraction of 1 mph.

Your program seems to be subtracting 0.3s from every trap time and adding 4 mph to every trap speed instead of just adding 1 foot to every trap distance.

I was premature with the Maltesers for you it seems.

[Pumaracing]

Just to be clear. Does YOUR data set include any rotating inertia at all? I thought it had none but could be wrong.

Also what does your data set have have for clutch slip rpm and peak rpm / gear change rpm?

[Pumaracing]

OK. I've had yet another go and think I've answered my own questions.

I don't think you've ever posted results from a single data set before with both roll-out and no roll-out so I assumed this must be something to do with it.

To get the 60 ft time so high I think you've used a very low clutch slip rpm, I match your time using about 2700 rpm. Then the car accelerates quicker again as the revs build up until it hits the grip limit you've used to try and match the 330 ft time and finally the 1/4 mile terminal speed is so high because you've created a power curve that exactly peaks out in 4th as the car hits the 1/4 mile and also because the gear change time is so quick.

If the above is correct then I think we've finally agreed the programming calculations.

[Pumaracing]

Stan Weiss said:

Dave,
Yes, I have it set at 2675 and much faster shift times than you do.

Stan

Luv a duck Stan. No wonder I had such a job trying to work out what you'd done. This may however deserve some sort of award for the most errrm "creative" juggling with the variables in a simulation to try and match a set of bum speeds and times ever seen.

As to real gear change times with manual boxes, look at the data Stevieturbo posted. You can see the width of the gaps in the acceleration curve at each change and it's about 0.5 seconds from power off to fully back on again. However some acceleration is still taking place as the tyres and clutch grip again so the total time lost is a bit less than this.

From simulations over the years checking car magazine road tests I reckon the best a professional test driver can do is about 0.25 to 0.3 seconds when it's not his own car and he doesn't care about breaking anything. It's not easy for an owner to be quite so brutal.

[Pumaracing]

I recall my mate in an F1 team telling me about the new seamless shift boxes that were coming in 7 or 8 years ago. It wasn't just that the change time was down to about 15 ms or so, maybe even less now, but somehow you actually gained a little kick of power by utilising the inertia of one gear cluster that was already spinning up but not yet engaged as another cluster disengaged.

It was all very hush hush and he couldn't go into details but I doubt it would have been comprehensible to mere mortals anyway. I have enough trouble remembering how a normal gearbox works.

The one thing I never built into my simulation which does have a small effect on very high speed vehicles is the loss of speed during the gear change which then has to be made up again. You can see about 2 mph on your own car on the 4th to 5th change.

[Pumaracing]

Stanley,

I have a little exercise for you. When an engine revs at full throttle in neutral the power is being absorbed solely by the engine's own internal component inertia. By timing the rate of increase in rpm from idle to rev limiter it should be possible to exactly calculate the shape of the entire torque curve.

With a known engine bhp the rotating inertia should be calculable or vice versa.

I assume that these ecu data loggers might be able to time the rpm increase. I'm struggling with the maths to program the power curve though. Reckon you can crack it?

[Pumaracing]

Well I've come up with an answer Stan although my head is hurting so I'm not 100% sure it's correct. By my calculations an engine producing 100 ft lbs of torque ought to be able to accelerate an inertia mass of 25 lbs at an average radius of gyration of 3.5 inches at about 14,500 rpm per second.

In other words to go from 1000 rpm to 6500 rpm should be about 0.4 seconds.

Does that look right to you?

[Pumaracing]

Excellent. I agree your calculations exactly at 0.3773 seconds with 100.9 ft lbs.

So now all we need is some data from people with ecu logging capability of engine acceleration rate from idle upwards to red line with engines of roughly known power curves. Then we can calculate the engine inertia properly.