OptimumLap, Lap Time Simulation help

[Flight909]

Hi all,

I'm planning a trip to UK later this fall to test some of the classic tracks with mine modified M3. I planning some track days at Brands Hatch, Snetterton and Cadwell Park.

To understand the tracks and help me to prepare, I used OptimumLap which is a free lap time simulator (www.optimumg.com/software/optimumlap). We have used it a lot in my Formula Student team. It is using a simplified vehicle model, so it is not to be used to do a detail setup of the car, but it is really quick and easy to use to get the "big picture" of the performance of a car on a given track. I used it recently to decide to upgrade the aerodynamics of my car instead of more power, because the simulation shows I would gain more with more downforce. It is also helpful to understand shift points and such things.

I have found Brands Hatch and Snetterton in their online track data base (share.optimumg.com/tracks), but I have not found Cadwell. Does someone have logged data for that track? They have an online tool which can be used to convert logged data into tracks in the software. I would really appreciate if someone could create that!

Here are some pics of the program if you don't know it:






Danke!
Florian

[Adam43]

Hello and welcome to 10-10ths.

I might be able to help. I have a gps racelogic performance box. I must have some Cadwell data. I haven't been for a couple of years, but I reckon I must have had the it then. You can use lateral g and speed and the helpful utility on their website to 'make' a track. I did this with Donington and it worked well. I'll have a look over the next couple of days.

I've only played with OptimumLap over one evening. It gave a realistic lap time over Donington. However the sim top speed was too high and the cornering speed too low compared with real life. The gear predictor was a off too, but that might be down to the car being a classic and this assuming you were using a quick sequential change. I couldn't correct without setting some unrealistic parameters. Still I did only have one evening with it and didn't read up on the best ways tweak the output.

I'll let you know about Cadwell.

[Flight909]

Originally Posted by Adam43

Hello and welcome to 10-10ths.

I'll let you know about Cadwell.

Adam thanks for this!

I was able to get fairly good matching between my apex speed in the simulation and in my car (DL-1). I think my apex speed all where acurate to +/- 5 km/h which I think is quite alright. Top speed on straights was a little bit high, but I think it is because I'm a slow shifter. I have mainly compared for Hockenheim.

I learn a lot from using this, helping me understand better my cars performance. It would fun to read what other people have learned for their cars. Maybe I can learned something more!

[Adam43]

I did have some Cadwell Park data and have made a track from it. Private Message me and I'll send you the file.

I wondered about shift time. I don't think it is enough to explain the top speed difference. The sim spends much longer at the apex speed that I do in real life. So I have a higher apex speed and less time at it.

I have some drag strip runs and I was going to model that to get the power/gearing right and then work on the apex speed.

[phoenix]

Just a comment on predicting/calculating speed on straights:

I know nothing, zip, about the software you are using, but to give accurate top speed/time-to-top-speed it will need to evaluate/predict drag very precisely too.

As aero drag is probably the least predictable element of any car (without a wind tunnel + straight line testing as F1 teams use) I am not surprised that erroneous values are output by OptimumLap.

Just my thoughts...

[tristancliffe]

OptimumG (been playing with it last night and this morning as a result of this thread) doesn't get it right. Despite lots of fiddling using as many 'knowns' as possible, and playing with the unknowns (sensibly), it is impossible to get the accurate outputs of top speed, min speed, lateral and long g at various speeds etc etc.

However, it isn't a bad comparison between different specs. So, if you want to know how much time you'll save (in terms of percent) by using a better engine, or fitting a lighter battery then it sort of quantifies it. A bit.

For instance, it suggests that 10kg is worth 0.25% (or 0.15 seconds per minute), or that another engine might be around 0.8% (or 0.5 seconds per minute) etc.

But of course, you can skew the data so that it shows what you WANT to see, rather than what is necessarily the reality of it. i.e. use with caution, and treat the results with suspicion.

[Adam43]

No problem, I have phd in Physics. Suspicion of modelling comes naturally.
I was impressed how close it was. I was also I impressed with how easily you can add new tracks. Never heard of the software to checking a laptime on a track I'd 'made' from GPS data in an hour. And eating a sandwich.

[greenamex2]

Downloaded it and spent a few hours playing. Not perfect but near enough. Putting in average tyre figures rather than peak it predicted a laptime within 3 seconds of an actual. A couple of seconds were down to car issues on the day so probably accurate to within a couple of seconds. Comparing speeds against actual they were pretty close.

As said earlier, it's biggest use is then changing things and estimating the effect on times. I didn't totally agree with the results but they did highlight some areas for further investigation and may turn out to be correct...if so then I have some big gains to come from the car!

Most importantly, it is really easy to use and FREE!

[Flight909]

Adam,

Thanks for the track - I found it in the trackdatabase! Thanks a lot for this!

About results accuracy, I think it is not bad with how simple it is. For example, with a more complicated software I would need to have Pacejka model for my tires, something I don't have (it is different for my Formula Student team). tristancliffe, if you have a down force car, maybe you must use the advanced tire model in OptimumLap to get more realistic effects. I know my Formula Student team did this because they have a light car and much down force, so you must have tire load sensitvity.

I also noticed that using the Hockenheim that is in the trackdatabase do not give me great results, I think because the track is made from data from a car that make a different drive line (maybe that data was from a FWD car). I created my own Hockenheim and I got better correlation.

I did not have too much problem with with finding the drag number, because I have made top speed runs with my car (thank you Autobahn) and I know pretty well what is my top speed. So I adjusted drag number until I got the correct top speed. The RPM for my top speed is about 150 RPM away from what I see in reality.

[phoenix]

I downloaded the program and have been playing with it. I could do with some more UK tracks - like Brands Indy, Castle Combe and I would find Goodwood useful too. So if anyone has any data that would be great!

[NJH]

I concur with Pheonix on the aero drag. I developed a model of a Porsche 968 race car in rFactor a couple of years back with Hartech (Porsche club championship) and had the benefit of knowing both detailed tech data on the real car and data logging from most UK tracks. The aero drag figure in their physics file I just had to keep tweaking until I could match the terminal speeds on the straights. Some of the numbers ended up a bit arbitrary but I didn't have much choice really as the aim was more to match a real cars performance than to generate a perfect physics model. Also be a bit careful with the corner speeds, I could match most corners accurately (same mph figure in whole units) but I always found that low speed corners gave unrealistically slow performance. I imagined this must be down to the tyre model but didn't have the time to start playing with that as this was something I was doing partly for fun and partly for experience in my spare time.

[Flight909]

I found this interesting article about OptimumLap. I can't even start to imagine what the F1 teams are using to simulate the car...

http://www.optimumg.com/technical/si...es-grand-prix/

[Notso Swift]

Predicted time 97.53 seconds
Actual time 1:35.657

When you consider the lack of real data they would have it probably isn't that bad, and the HRT's would have loved to have that time as it would put you 11th on the grid.

I actually love the software as I can do a little bit of fantasy work on my car and see haw fast I could be

[salva]

I'm using Optimum Lap for time simution of our FSAE race car in an acceleration test. We intent to select the final drive ratio based on this analysis, however we're having a problem, the results implied that the shorter our final drive ratio the fastest our car is gonna be independently how short it is and the graphs shows a linear growth that seems incorrect by the physics analysis of the problem. Therefore we know that we\'re inserting some data in the wrong way.