RomRaider

Once the polynomial that is generated to fit the raw RPM vs Time data, you could just take the coefficients and produce the derivative (acceleration), right? You could then generate the Torque/HP curves without referencing the raw data any more (I think). Is this how the final curve is generated?

What do the dots represent?

Currently I just calculate slope between two points to find acceleration... but the derivative would be very easy to do, as yes, I have the polynomial coefficients. Let me check that method out.

Dots are no long used and have been removed.

Now the long explanation...

1) Time and RPM are collected into a series and from this dataset a least squares polynomial is derived.
2) Smoothed RPM data is calculated using the polynomial coefficients and the original time range but stepped through at 0.01 second intervals.
3) Power and torque are calculated using RPM acceleration (derived from the first derivative of the RPM polynomial) and mass. The power and torque values are stored in two new series against RPM and plotted. Peak power and torque are identified and area under the curves is calculated.

Changing the smoothing value (which is the degree of the polynomial) affects the original RPM smoothing and therefore the power and torque curves once recalculated.

Dale

EDIT: Updated steps above

Just to say this feature is absolutely fantastic! My result showed 385 whp which is about what I expected with my set up. Thanks guys for this excellent piece of work!

That's great news, thanks for the feedback.

After step 1) you have needed polynomial factors.
Just recalculate them into torque and power polynomial factors analytically and draw the curves (vs time) if needed.

There is absolutely no need to recalc a smoothed RPM curve relative original time stamps. Just use polynomial order+1 points.

After that you may recalc power and torque vs RPM polynomial factors if those curves are needed. You can save a lot of time in such a case because just 10-15 initial "interpolated" points are really needed.

Thanks. I've been fine tuning the calculations and working in the use of the first derivative for acceleration.

New test version rd-v10 attached to post topic39-34.html

- There are now comparison statistics between the current trace and the Reference Trace for the time values (sorry no power and torque at this time).
- I've introduced a new statistic; Area Under the Curve (AUC). This value is derived from:
(power*RPM + torque*RPM) / (time from 3k to 6k RPM) / (A constant I chose)
It's a subjective number, only used for comparisons and will be in the range of 50 to 250.

This version has no graph dots to represent power and torque raw values. Interpolation of the RPM is done then power and torque calculated directly and then plotted, no more in between step. This should provide more consistent results as you change the smoothing. Adjust the RPM range as needed if the curve shows your power and torque peaks are off the right hand side of the graph as it may shoot up at times, 2000 to 6500 works best.

Hi a couple of requests if I may:

Would it be possible to select EDM as well as USDM and ADM cars?

Also here in the UK BHP is measured rather than WHP so would it be possible to have this BHP as an output option?

Thanks again

Alistair

If you can provide EDM model information in THISthread, in the correct format, Dale will be able to add it to the next release.



BHP and WHP is on and the same - unless BHP is really refering to crank HP, in which case it is not possible to accurately calculate crank horsepower without first knowing driveline losses.

Leslie

thanks will do re the EDM definitions. Where did you get your data for the ADM cars just out of interest? Just thinking about how to obtain info on frontal area etc.

Yes in the UK BHP is generally considered to be crank HP. I believe delta dash spits out BHP from the graphs I have seen so would be great if we could do the same with this tool. I wonder if there is any data out there for 5mt and 6mt losses which could be incorporated?

I was considering BHP but didn't look too hard on the InterWeb for any info on the driveline losses.
I’ve seen these parameters used:
1st Gear Loss 15
2nd Gear Loss 10
3rd Gear Loss 7
4th Gear Loss 3
5th Gear Loss 0
6th Gear Loss 6
Where the loss is based on the gear ratio I believe. I assume this is % of loss and 0 is at the 1.000 ratio but not completely sure.

As for the car defs, look at the XML file in a text editor and it's fairly clear what is needed. If you have the info in a spreadsheet I can easily convert it.
The frontal area is defined to be the same for all cars. Going +/- a square foot will not affect your results much more than 1 or 2 HP overall.
The biggest factors are (in order of impact to the final result (lower is better in every way)):
Weight
Elevation
Temperature
Frontal Area
Humidity

thanks - surely there would need to be some transitional loss in all gears due to friction along the drivetrain?

Jeff Sponagle had a very interesting thread a few years ago on NASIOC about gear loss. it wasn't near what some people believe.

This is a very nice addition, i tried it a couple weeks ago when it was snowing alittle bit (bad idea). check out my graph


i think it was about 5000 rpm when i hit the slick spot lol. Thank jeebus my ass end didnt shoot out when it happened. oh well lesson learned i guess.

thanks again for everyones hardwork.

oh do you still need someone to try it on a automatic? i got either a 08 legacy gt or a 09 forester xt if you want me to i would love to help.

Oh yes an automatic plot would be great if you could, and maybe post your romraider.log file from the install directory afterward too (zipped of course).

But first, grab the newest version 10 from this post.
topic39-34.html
Thanks.