Powercalc vs ebike.ca simulator

[Busybee]

I recently became aware of a cycle power calculator, Powercalc, shareware with an obnoxious nag screen, from http://www.machinehead-software.co.uk/cycling.html

It is for bicyclists but includes power in watts required under specified conditions. The ebike.ca simulator has to calculate the load wattage which is independent of power source, and so the two simulators should give similar results. The Powercalc program, however, does not agree with the ebike.ca simulator especially on zero slope comparisons. This is of concern because claims by some ebike manufacturers may be based on Powercalc estimates.

Assuming the Powercalc default of an 85 km (187 lbs) bike/rider on a race bike in a tuck position using 700c wheels, the two simulators are in reasonable agreement assuming a 10% slope (Powercalc being 1% to 12% lower for speeds from 5 mph to 30 mph respectively), and in lessor agreement assuming a 5% slope (Powercalc being 5% to 18% lower for speeds from 5 mph to 30 mph).

The big difference however was at 0% slope. Powercalc was 36% to 44% lower, but unlike on slopes, the Powercalc estimate was lowest assuming lower speed. This is bad as most ebike claims are based on travel on flatland. The only certainty is that both simulators can't be correct.

If some ebike manufacturers are using the known output of their motor and using Powercalc to estimate speed, then claimed range and speed would be quite a bit more than the ebike.ca simulator would predict.

Since ebike.ca also has plenty of Cycle Analysts to check their simulator against real world performance, I'll bet Powercalc is the one more in error. Still, could explain where some of the inflated claims may come from.

 

[Russell]

To check if the calculators are in agreement you need to make sure the vehicle parameters are identical, especially the CdA and Cr values.

My personal favorite regular bike power calculator is this one: http://www.kreuzotter.de/english/espeed.htm


-R

 

[Busybee]

Have used the kreuzotter site and just run all the same numbers. Results are in between. At 10% slope kreuzotter is 1% to 5% lower than ebike.ca. At 5% it is 5% to 9% lower. At zero slope it is 21% to 37% lower but like Powercalc it differs more at lower speed. At least kreuzotter shows the math. As kreuzotter is similar to Powercalc, just not as low compared to ebike.ca, that suggests ebike.ca values for watts required are on the high side especially at zero slope.

 

[teklektik]

Have used the kreuzotter site and just run all the same numbers. Results are in between.
At 10% slope kreuzotter is 1% to 5% lower than ebike.ca.
At 5% it is 5% to 9% lower.
At zero slope it is 21% to 37% lower but like Powercalc it differs more at lower speed.
...
As kreuzotter is similar to Powercalc, just not as low compared to ebike.ca, that suggests ebike.ca values for watts required are on the high side especially at zero slope.

Some time ago, I verified the math for the ebikes.ca Load Line calculation in terms of weight, slope, and drag. This was a verification of calculations based on the physics, not a comparison with other calculators.

• I have no doubt that the Load Line calculation is correct and the plot is displayed accurately.

Beyond the math verification, the simulator has historically been shown to give good estimates that can be verified empirically - unlike the bicycle rider world where instrumentation is problematic. The reported discrepancy of 37% on zero slope is huge and certainly would be easily seen by any ebiker who has compared estimated and actual performance and range. It's a matter of speculation as to why the other calculators vary so widely from the (proven) ebikes.ca simulator.

• That said, the ebikes.ca simulator does have a minor glitch in that it uses at air density at 5 degC (44 degF) instead of an arguably more reasonable temperature of 25 degC (77 degF). This causes slightly elevated aero drag (7%) because of denser air at lower temperatures. Because there are other terms in the calculation, the resulting net error is reduced - even more so as slope is increased.
  
  The plot below shows the effect for a the default ebikes.ca Mtn bike with 100kg gross weight. The net motor power discrepancy due to temperature at 30mph is 60W.
  
  
  Justin has this on his ToDo list when the Simulator gets the next round of upgrades.

 

[Busybee]

are on the high side especially at zero slope.

Oops. Should have said "may be on the high side" instead. I plan to continue to use the ebike.ca simulator and discount the other two.

 

[Drunkskunk]

My understanding on the main difference between Ebikes.ca simulator and the various bike calculators is that the Ebikes.ca simulator is based on actual measured motor performance on a dyno, where the other calcs are based on theoretical calculations, but no real world testing.

 

[teklektik]

Actually, Justin's simulator has two parts - the Load Line calculator and the motor simulator. The first is motor-independent and computes the power to propel the bike based on drag coefficient/frontal area, rolling resistance, weight, and grade - pure physics without any motor dependencies. The second part operates as you mention - using data derived from dyno runs to drive a battery/controller/motor simulation. The intersection of these two curves - the (black) Load Line from part one and the (red) motor power curve from part two - is the expected speed (where the power needed to propel the bike equals the available motor power).

So - here we were talking about the Load Line part which is pretty standard stuff and should be exactly the same as bike calculators. Some differences in results may arise because of hidden inclusion of drive line losses or differing aero and rolling coefficients for the same claimed bike type.