[steamer_ctn]

Hi all,

As part of the work on the steam module, I decided to run my own "Rainhill Test". The purpose of the test was to see how the steam module would cope with an extremely lightly powered locomotive (Tractive effort of approximately 820lbf) A mere fraction of modern steam locomotives.

The original Rocket, by Joseph Spinella, is available from uktrainsim. The original model is set up to work in MSTS, and it appears to have been necessary to power the locomotive at more then its rated tractive effort in order to get it to move.

Testing with the new OR module, and lower power levels, also demonstrated a reluctance for the locomotive to move, and based upon research I believe that the current representation of friction in OR, which was based upon MSTS is not realistic, and especially the static (starting) friction which appears to be in excess of agreed railway industry figures. Most of my research has been based upon the information in the Baldwin Locomotive Data book - 1944.

As a result, I have made some modifications to the OR friction model based on this publication as follows:
i) Set the journal and roller bearing friction values at the values suggested
ii) Set different operation of roller and friction bearings between 5 and 35mph as described.
ii) Enabled OR to read the Davis values (as calculated with FCalc 2.0)

These changes will apply for all rolling stock and only come into effect if the new Davis values are included in the ENG or WAG file. If the Davis values are not present in the files then OR will default to the previous implementation of friction.

One way to get an appreciation of differences, if experimenting, is to watch the "HUD - Force Information" screen to see how the "Friction" value is changing.

The changes implemented model the differences for friction and roller bearings internally in the program, based upon the friction bearing as a reference, so the Davis values should be those calculated for friction bearings.

To set your stock up to take advantage of this you will need to add the following new OR parameters.

Wagon section of ENG file or WAG file.

Comment ( ********* Start OR Steam Module Parameters - Wagon Section ********* )	
	Comment ( Speed: 100km/h, Weight - 4.704 tons (US), Drv Weight - 2.191 tons (US), Drag - 1, Area - 2.5m2, Bearings - Friction, Axles - 2 (1 Drv, 1 aux), Type - Steam Locomotive Standard )
	
	ORTSDavis_A ( 502.8 )
	ORTSDavis_B ( 1.5465 )
	ORTSDavis_C ( 1.437525 )
 Comment ( ORTSBearingType ( Roller ) Comment ( If friction bearing, then leave out) )
	
Comment ( ********* Finish OR Steam Module Parameters ********* ) 

Engine section of ENG file

In addition to other OR parameters add this one, if appropriate:

ORTSDriveWheelWeight ( 2.12t ) Comment (If not steam locomotive, then leave out)

These changes have been implemented with #1978.

Cheers

Peter

[cjakeman]

Hi Peter,

steamer_ctn, on 31 January 2014 - 10:07 PM, said:

	ORTSDavis_A ( 502.8 )
	ORTSDavis_B ( 1.5465 )
	ORTSDavis_C ( 1.437525 )
	ORTSBearingType ( Roller )
	ORTSDriveWheelWeight ( 2.12t )

Are you able to add these new parameters to the Operations Manual (\Documentation\Manual.doc) in the appendix ORTS Parameters ?

[Genma Saotome]

What are the A, B, and C parameters?

Will it all work if the driving wheel weight is missing... or do you mean locomotive weight on the driving wheels?

[steamer_ctn]

Hi Chris,

cjakeman, on 01 February 2014 - 05:30 AM, said:

Are you able to add these new parameters to the Operations Manual (\Documentation\Manual.doc) in the appendix ORTS Parameters ?

I will have a look at it. I will need to add the other parameters that have been created for the general steam model.

Hi Dave,

Genma Saotome, on 01 February 2014 - 09:33 AM, said:

What are the A, B, and C parameters?

Will it all work if the driving wheel weight is missing... or do you mean locomotive weight on the driving wheels?

I not sure what info you want for the first question. They are the standard Davis values.

In regards to your second question, it will still work if the "locomotive weight on the driving wheels" is missing. It uses an assumed factor of adhesion to calculate the weight.

Cheers

Peter

[Lindsayts]

Genma Saotome, on 01 February 2014 - 09:33 AM, said:

What are the A, B, and C parameters?

Will it all work if the driving wheel weight is missing... or do you mean locomotive weight on the driving wheels?

The davis equation is a standard 2nd order quadratic equation, these take the form of "ax squared + bx + c = y". In the case of davis (a railway engineer from the 1930's) these equations are used empirically as it was found the curve produced by the equation closely matches the resistance force curve of railway rolling stock. The values a,b and c are chosen empirically, ie by experimention so the curve produced fits the resistance curve of the rolling stock concerned.

Different railway administrations have used varying values for these to fit the particular circumstances. The ones published in the "Fcalc" program are just some of several differing values.

I use a mathematical graphing program to adjust these values so the curve makes a "best" fit for the rolling stock concerned.

Lindsay

[Genma Saotome]

steamer_ctn, on 01 February 2014 - 11:49 AM, said:

Hi Dave,

I not sure what info you want for the first question. They are the standard Davis values.

In regards to your second question, it will still work if the "locomotive weight on the driving wheels" is missing. It uses an assumed factor of adhesion to calculate the weight.

Cheers

Peter

What I mean is the Davis Formula that I know takes the form of "R = 0.6 + 20/w +.01V +KV^2/wn" (one example) and I don't see any A, B, or C there. If, perchance what I wrote is what you mean by R = 0.6 + A + B + C then may I suggest the issue is why is the end user being asked to do any of the math?

IOW, anything the end user is expected to provide that is named, essentially, A, B, and C might be done better by having parameters of NumberOfAxles() and CrossSectionArea(), stuff they might know how to provide, and leave the math to the program.

Also... what happens to A, B, and C if total weight gets relocated out of the .wag into the consist file?

[copperpen]

The formula R = 0.6 + 20/w +.01V +KV^2/wn is merely an adaptation of the base Davis equation and is R=0.6+A+B+C. The three items A, B and C are derived from the Fcalc output which most will have in the eng or wag files, and correspond to the first, third and fifth items of the MSTS friction line. There is no need for anyone to add any of these ORTS prefixed lines to their eng or wag files, OR will happily run without them. No matter which is used, the program is doing the math, not the user.

As to what happens if the things get moved, just move the link to where the data actually resides.

[Genma Saotome]

I guess I'm really confused Mervyn... sorry for all of the questions... first: are you saying people have to run Fcalc in order to obtain values for A, B, and C? And/or are you saying people have to transcribe values out of the existing MSTS Friction line... most of which are bogus?

Second, my major concern is that it appears that you are asking the end user to provide values that are the result of some computation, suggested by the use of the code letters A, B, and C, instead of asking them to provide the source values that will be used by some computation. IOW, it appears that rather than ask the end user to enter a value for, say, CrossSectionArea() or ActualNumberOfAxles() parameters whose names convey their meaning and therefore are easily understood by all you are asking people to provide ORTSDavisA() for which, at least in this thread, has not been described as to what it is nor does its name suggest anything. Saying, oh, it's a standard element of the Davis formula does not explain to Joe EndUser what to enter. I think that is a serious problem.

[steamer_ctn]

Hi Dave (and others who I might have confused),

Firstly, please refer to the Baldwin documentation that was in the first post, and the word document that is provided with FCalc, for background information.

Genma Saotome, on 01 February 2014 - 01:56 PM, said:

What I mean is the Davis Formula that I know takes the form of "R = 0.6 + 20/w +.01V +KV^2/wn" (one example) and I don't see any A, B, or C there. If, perchance what I wrote is what you mean by R = 0.6 + A + B + C

The Davis formula for a steam locomotive is, as described in the FCalc document, (which can also be aligned with the Baldwin Document formulas):

R=1.3T + 29N + 0.03TV + 0.0024CAV² + 20DF

The ABC values are found as follows from the above formula, when the formula is laid out in this form - A + BV + CV^2:

A = journal + mechanical friction = 1.3T + 29N + 20DF (Note 20DF is not present in diesel or electric formula)
B = flange resistance = 0.03T (varies for other types of vehicles)
C = air resistance = 0.0024CA

FCalc calculates the Davis values and presents them either in terms of MSTS values, or the equivalent ABC Davis values. (see attached screen shot)

Genma Saotome, on 01 February 2014 - 01:56 PM, said:

then may I suggest the issue is why is the end user being asked to do any of the math?

IOW, anything the end user is expected to provide that is named, essentially, A, B, and C might be done better by having parameters of NumberOfAxles() and CrossSectionArea(), stuff they might know how to provide, and leave the math to the program.

The reason that the user is asked to calculate the ABC values is as follows:
i) From a pragmatic point of view, there are a significant number of different formulas for each of the different types of vehicles (as seen in the FCalc appendix), and the effort to include these variations was considered beyond the current scope, especially as
ii) FCalc already provides a user friendly GUI for the user to make the appropriate calculations as a relatively easy process
iii) Some users like to have the ability to make some level of cutomisation, the ability to enter the ABC values supports this, whereas having OR make the calculations will remove some of this freedom.

Now that these values are set up, future development could be undertaken to allow OR to do more of the calculations.

Genma Saotome, on 01 February 2014 - 01:56 PM, said:

Also... what happens to A, B, and C if total weight gets relocated out of the .wag into the consist file?

This will be an issue regardless of whether the new ABC values are implemented, or the existing MSTS style approach is retained. If dynamic loading is adopted then OR will need to be developed to do suitable back end calculations to pass the ABC values to the friction module.

Please let me know if there is additional information that I can provide and document, over time, to assist users in understanding the ABC values.

Cheers

Peter

Attached thumbnail(s)

• 

[copperpen]

Dave

The current situation with MSTS is that to get away from the default Kuju friction figures, a person has to use Fcalc and then put those numbers into the friction line. With the new Davis calculation added by Peter the end user has two options. First is do nothing and OR will work with whatever is in the eng/wag file. Second is to use the new ORTS specific lines, run Fcalc to obtain the Davis figures and put those into the appropriate A, B and C lines. Alternatively just use the first, third and fifth lines from the friction lines as the A, B and C. If one really wants to run realistically, there is also the train forces calculator by Al Krug to work out how much HP you need to move the train at X speed over Y grade.

I can see a problem ahead with activities if one person makes an activity using one set of resistance and the end user is using a different set.