Whats a good way to assign maximum score time in a lap sim?

[Spetsnazos]

So the next step has come in the lap sim and I would like to relate variables to points in the competition. The lapsim outputs time that is VERY close to the time that uni's score at the actual competitions.

I was thinking of taking the fastest time in each of the events at the competition and making that the maximum score, and from there have the software output the score based on those results.

Is there a better way of doing this? Which competition should I take the fastest times from?

Goal is to follow Big Bird's thread.

Cheers.

[Spetsnazos]

So the next step has come in the lap sim and I would like to relate variables to points in the competition. The lapsim outputs time that is VERY close to the time that uni's score at the actual competitions.

I was thinking of taking the fastest time in each of the events at the competition and making that the maximum score, and from there have the software output the score based on those results.

Is there a better way of doing this? Which competition should I take the fastest times from?

Goal is to follow Big Bird's thread.

Cheers.

[Tom W]

Whichever competition you took the track from?

[exFSAE]

Just curious... do you know if your simulation is producing directionally accurate results? E.g., if you make changes X, Y, and Z on the car that make it turn faster or slower laps... does the sim correctly rank them?

IMO getting the absolute lap time correct isn't that important, or isn't as important anyway compared to finding directional responses.

Without having that proven, there's a danger in chasing these sim results down the wrong path.

[Spetsnazos]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by exFSAE:
Just curious... do you know if your simulation is producing directionally accurate results? E.g., if you make changes X, Y, and Z on the car that make it turn faster or slower laps... does the sim correctly rank them?

IMO getting the absolute lap time correct isn't that important, or isn't as important anyway compared to finding directional responses.

Without having that proven, there's a danger in chasing these sim results down the wrong path. </div></BLOCKQUOTE>

Changing X, Y, Z = changing vehicle weight, vehicle parameters, etc, yes it affects lap time in the correct direction.

I meet all of the above criteria, now should I get acceleration/skidpad times from the same competition as I used the autocross/endurance course from?

[Spetsnazos]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Tom W:
Whichever competition you took the track from? </div></BLOCKQUOTE>
That was the plan for now.

Anyone have any more input? Thank!

[exFSAE]

Acceleration event might be one of the more difficult things to simulate.

[Spetsnazos]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by exFSAE:
Acceleration event might be one of the more difficult things to simulate. </div></BLOCKQUOTE>

not sure if serious.jpg

[exFSAE]

I am absolutely serious. So much of the ET is how you get off the line... how the tires hook up from a standing start.

The tire dynamics are a big challenge in that type of situation... not to mention all the drivetrain inertial effects, revving up and dropping or half slipping the clutch.

If it were me I wouldn't spend too much time on it. Better things to do with those hours.

[Spetsnazos]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by exFSAE:
I am absolutely serious. So much of the ET is how you get off the line... how the tires hook up from a standing start.

The tire dynamics are a big challenge in that type of situation... not to mention all the drivetrain inertial effects, revving up and dropping or half slipping the clutch.

If it were me I wouldn't spend too much time on it. Better things to do with those hours. </div></BLOCKQUOTE>

True, which is why I am not expecting exact real world results. Things like weight, power, tires, cg, etc will have an affect on acceleration and its nice to know how much of an effect.

Obviously with a sim there are assumptions and approximations. As long as its reasonable of course...

[nowhere fast]

You can run the sim for all of the design concepts you are interested in, or expect to see in competition and take lowest time achieved by any concept in each event as the Tmin.

If you take Tmin from an actual competition then your results will be more dependent on the absolute accuracy of your sim, but if you let the sim find Tmin then you should get decent results provided the sim is directionally accurate.

[Spetsnazos]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by nowhere fast:
You can run the sim for all of the design concepts you are interested in, or expect to see in competition and take lowest time achieved by any concept in each event as the Tmin.

If you take Tmin from an actual competition then your results will be more dependent on the absolute accuracy of your sim, but if you let the sim find Tmin then you should get decent results provided the sim is directionally accurate. </div></BLOCKQUOTE>

Great point, I like this the best. I'll have to run some tests and determine the optimal set-up to get Tmin. Going with this option http://fsae.com/groupee_common/emoticons/icon_smile.gif

Cheers thanks!

[StevenWebb]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by nowhere fast:
You can run the sim for all of the design concepts you are interested in, or expect to see in competition and take lowest time achieved by any concept in each event as the Tmin.

If you take Tmin from an actual competition then your results will be more dependent on the absolute accuracy of your sim, but if you let the sim find Tmin then you should get decent results provided the sim is directionally accurate. </div></BLOCKQUOTE>

this is the only way to do it.

You'll also get into the problem of how many points to get. If your sim is aiming for maximum points then you could end up with 900 points but the team coming second might end up with 899 points. On our (Monash's)competition simulations we compare our car points potential to the average of the other simulated cars as well as the highest scoring competitor.

[exFSAE]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Spetsnazos:
Obviously with a sim there are assumptions and approximations. As long as its reasonable of course... </div></BLOCKQUOTE>

Again, this comes down to proving (to a design judge for example) that your simulation is yielding reasonable answers.

[Pete Marsh]

T min is not critical to the relative difference between results.

Pete

[Spetsnazos]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Pete Marsh:
T min is not critical to the relative difference between results.

Pete </div></BLOCKQUOTE>

Sure it is, its great for determining how many points you are gaining/losing. 1s =/= 1 point

[Pete Marsh]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content"> Sure it is, its great for determining how many points you are gaining/losing. 1s =/= 1 point </div></BLOCKQUOTE>

Yeah, regardless of what value you choose for T min. My point exactly. The event rules call for 1320sec, just use that.