[scottb613]

Hi Folks,

Working on optimizing some of my projects for OR...

Question - I found an example where we include "Davis Coefficients" with the friction settings in ENG/WAG files... I'm assuming we should include the ORTSDavis_A, ORTSDavis_B, and ORTSDavis_C values in both ENG/WAG files to improve OR performance ? I'm also assuming the vehicle of choice to calculate friction values is still FCALC2 ? So - with FCALC2 I noticed that the Davis Coefficient values change dramatically when switching between US and Metric units - however - no unit of measure is included in FCALC2 - in addition - the example I found of an ENG file that includes these values also does not include a unit of measure...

We're probably using the "Metric" setting and no unit is required ?

While FCALC grabs the speed from the ENG file for locomotives - the WAG files do not have a value for this - should I leave the speed set to the default "0" in FCALC2 for proper results with WAG/Freight Cars ?

Any help - much appreciated...

Regards,
Scott

[Lindsayts]

The davis co-effiecents are dimensionless factors (I think), I have always used imperial measures as these produce realistic data when the davis equations are applied by hand. Metric data being of course converted to imperial. I have not done much with the metric davis equations as they do not appear to produce good data (I only tried once though so I could have mucked it up).

I have done much work though with the imperial davis equations given with the Fcalc2 docs (there are other sources) mostly with passenger vehicles including DMU's, results equate well with reality.

Lindsay

[scottb613]

Hi Lindsay,

Thanks for responding - yeah - I think I mucked it up as well... My test is a new very large Consolidation - similar to the Western Maryland H9's except mine have that extra wide firebox for Anthracite - I think I have the physics on the locomotive and tender - plausible (my very first physics setup on my own)- so I tried using FCALC2 to set up (40) 67 ton hoppers on level track... Before FCALC2 and the respective Davis values - I could start the train fine - after no dice... What started me down this road - is it seemed there wasn't quite enough resistance on a the test train - as even when going up a moderate grade - if I let off the throttle - it would coast for ever - nothing scientific - just the seat of my pants feel...

I'll revisit...

Do you think these David values should be included - or - just stick to the MSTS values ?

I was using Bob's "Plainsman" car physics to start with...

Regards,
Scott

[copperpen]

I have a 40 car test train set up using the ORTSDavis lines. The test car is the US2graincar set up as a 100 tonner. When I used FCalc to set these up it produced MSTS friction figures that were actually lower than those provided in the wag file. However, since I no longer use MSTS that was of no interest.

I use this set to test steam traction, in particular starting TE. I don't have a large WM 2-8-0, but I do have the old consol to play with. Two of those in top'n tail will shift my 40 cars with no problems.

My ORTS friction for this car is
ORTSDavis_A ( 222.5 )
ORTSDavis_B ( 1.5 )
ORTSDavis_C ( 0.057073 )
ORTSBearingType ( Roller )

That gives me a starting friction figure of 7811 for each car.

[steamer_ctn]

scottb613, on 17 December 2014 - 05:15 AM, said:

I'm assuming we should include the ORTSDavis_A, ORTSDavis_B, and ORTSDavis_C values in both ENG/WAG files to improve OR performance ?

All vehicles should have "friction" properties assigned to them for realism.

scottb613, on 17 December 2014 - 05:15 AM, said:

I'm also assuming the vehicle of choice to calculate friction values is still FCALC2 ?

That is a personal choice, FCalc makes it easier to calculate, but you could also hand calc them if you want.

scottb613, on 17 December 2014 - 05:15 AM, said:

So - with FCALC2 I noticed that the Davis Coefficient values change dramatically when switching between US and Metric units - however - no unit of measure is included in FCALC2 - in addition - the example I found of an ENG file that includes these values also does not include a unit of measure...

We're probably using the "Metric" setting and no unit is required ?

By default OR assumes metric units, if imperial units are used then, you would need to enter the imperial units into the Davis parameters in the ENG file.

FCalc uses metric units as input and provides metric units as output. Note: I believe that metric tonnes are used as input.

Note also that the metric and imperial formulas are different. See the FCalc document for example metric formulas. See the reference on the bottom of my friction page for an example imperial resistance (friction) formula.

scottb613, on 17 December 2014 - 05:15 AM, said:

While FCALC grabs the speed from the ENG file for locomotives - the WAG files do not have a value for this - should I leave the speed set to the default "0" in FCALC2 for proper results with WAG/Freight Cars ?

See the FCalc documentation for a discussion on the impact of speed, based upon my interpretation, any speed under 60mph is probably not worth trying to set.

[scottb613]

Hi Folks,

Thanks for the additional feedback - LOL - calculating the values by hand is probably beyond the scope of what I'm willing to do...
:I-Agree:

Ahh - "Metric Tons" - I've been calculating "Short Tons" where I needed a value - I'll correct that...

I'll try your values as well Mervyn - thanks...

Let me play with this some more...

Appreciate the help...
:oldstry:

Regards,
Scott

[Matej Pacha]

Just a note from the code:

FrictionForceN = DavisAN + FrictionSpeedMpS * (DavisBNSpM + FrictionSpeedMpS * DavisCNSSpMM);

That means you need to adjust your Davis coefficients determined for meters per second. Since Davis formula is an empiric equation based on measurements usually in some general speed unit such as mph or kmph, this adjustment is necessary. Eg., if your DavisB component is 1000 and if it was measured with km/h, you should use 3600 instead to keep the result in Newtons at the same level.
One more thing: Davis formula has many forms, e.g.:
ForceN = DavisAN + SpeedMpS * DavisBNSpM + SpeedMpS * SpeedMpS * DavisCNSSpMM
Result would be very different. If you are using these values, just be sure about what you are using.

King regards,
Matej

[Genma Saotome]

We should be recording the inputs to the formula and letting the code do it's normal thing while parsing before passing the "official, standardized numbers" over to a set of formulas. That's how just about everything else is done. It makes no sense at all to do this differently.