[gpz]

steamer_ctn, on 24 April 2014 - 05:43 PM, said:

(...) eventually achieve (...) additional research, thinking and investigation (...) a little while

Have you realized, that I also described a possible method to reach it above my sentence you quoted (through 2 posts), not just dropped in a vague idea? ;)

The options ORTSDavis_A, ORTSDavis_B and ORTSDavis_C must be considered as absolutely temporary options, should never be advertised to public as a general solution. Moreover, handling of the "C" values of FCalc is also defective principally. (Which is not an unsolvable problem for a software in development, of course, just everyone must keep it in mind.)

[copperpen]

While I agree in principle that an automatic calculation of the friction factors is a good idea, I totally dislike the point of removing the ORTSDavis lines from use. I prefer to have the ability to fine tune the resistance figures to match a known set of data. Therefore the approach to this for me would be to adopt the following method.

Automatic calculation of resistance to be over-ridden by the ORTSDavis lines if present in a wag or eng file, or to be made an option of automatic or manual calculation.

[gpz]

You are right, it is always a good thing if the values can be fine-tuned. There might be also a case when the exact values are available from an outside source, and in this case they should be possible to be entered.

Please note the "C" values for locomotives outputted by FCalc apply only for the leading unit, thus is unusable as an ORTSDavis_C value. In the future this should be entered from a "Passenger Car" type recalculation by FCalc, and the value should be multiplied by OpenRails code (by 7.06). (Sorry for always repeating it, it is kind of a "Ceterum censeo Carthaginem esse delendam" effect by me. ;) )

[copperpen]

I do not think that the use of passenger car calculation for a non lead unit is correct, unless that unit is close coupled and in in effect streamlines. This does not apply in a lot of cases, specifically thinking of US style locomotives excepting the Carbody A and B units.

[gpz]

It is pure logic. If you look at the table in file "FCalc 2.0 Variable Parameters.txt" (I quoted a part of it in one of my previous posts), you can see that the "C" formula for the locomotive is the same as for leading railcar. It means the middle locomotive has to use the same formula as middle railcar in the file, which appears to be the same as the formula for a passenger car.

Certainly every time a vehicle has a different shape than a box, a modification factor could be applied. Also, when cars with different sized shapes following each other an additional correction factor could be applied. But there is no information about this in "FCalc 2.0 Variable Parameters.txt", so I think FCalc didn't calculated with these. I think this might be just a minor difference compared to the multiplication factor of 7 between the leading and middle-unit formula.

[copperpen]

Fcalc certainly does take into account the difference between different types of car, however, you must accept that Fcalc was developed to generate friction figures for MSTS and therefore the output will definitely be off by a certain amount.

If you can develop in ingame method of applying these figures which then allows my test train to match a proven friction curve I will be convinced that what you propose is right. Until then I remain firmly in the ORTSDavis camp. :oldstry:

[jorgen]

steamer_ctn, on 20 April 2014 - 09:44 PM, said:

Hi,
I have made some minor code changes to the starting friction, which are included in #2191.
It now caters for three types of bearings - Friction, Roller, Low (Low Torque) -
Peter

Hi
At last, the wagons now roll as they should. (as they did in Or 0.9)

Did some test. 51 Ton test wagons with these lines.
Friction (C1) 580N/m/s (E1)-0.10 (V2)1.7mph (C2)4.201N/m/s (E2) 1.5

1. Default ................No row in wag/eng files
22 * wagon Weight =lbs (0-5 mph) =5000 Newton Friction, Then C1 value and raising with higher speed

2. Solid Bearings ..... ORTSBearingType ( Friction )
22 * wagon Weight =lbs (0-5 mph) =5000 Newton Friction Then C1 value and raising

3. Roller bearings .... ORTSBearingType ( Roller )
9 * wagon Weight =lbs (0-5 mph) =2043 Newton Friction Then C1 value and raising

4. Low (Low Torque) ORTSBearingType ( Low )
4.4 * wagon Weight =lbs (0-5 mph) =963 Newton Friction Then C1 value and raising
0 mph= 963N 1 mph=850N 2 Mph=770N 3 Mph=700N 4 Mph=650 5 Mph=580N
------------------------------------------------------------------------------
Here was the original Msts values
Friction 580N/m/s -0.10 1.7mph 4.201N/m/s
5.1N/rad/s 1 -1rad/s 0 1 )
ORTSBearingType (Friction Roller Low) (Not in test 1)
-----------------------------------------------------
Then I replaced it with Davis values
ORTSDavis_A (188.165)
ORTSDavis_B (2.49802)
ORTSDavis_C (0.045)
ORTSBearingType (Friction Roller Low)
Done test 2-4

The results up to 5 Mph was same, but the difference were then the train come up to 5 mph.
If using (MSTS values) the friction were 580 Newton
Using (Davis values standard freight) the friction were 188 Newton.
------------------------------------------------------------------------------

Facts about starting friction.Canadian National Railway Resistance Formulae
The starting resistance of roller bearings is essentially the same as they are in motion.
In general a resistance of 5 lb/ton or less should be satisfactory for roller bearing equipment

I'm as happy with the last change ( Low)

But I do not agree with, why it is between 0-5 mph as it gets higher friction.
Solid bearings and roller bearings are the major differences in construction.
It's cylinders that roll in roller bearings and solid bearings is the axis on a bed of wood and oil or something like that. Large differences.

Here it should be roller bearings having higher friction just between 0-0.5(1) mph then C1 value.

Another big thing. Default is Solid bearings (Train almost glued within the track in the start)
We have to add this line ORTSBearingType (Low) in every eng/wag files who have roller bearings.
But I have over 1000 wag / eng files.
Should I now add these lines in all these files it will take 17 hours for me.....

This must be done with Low as default.

When version OR1 comes out, will the public to run this fine simulator.
If we don't change this to (Low) I think public will be very disappointed at how the train behaves at start up that they even abandon the simulator.
Not everyone is as interested as we are.

Besides, it is very difficult to find the specific eng and wag file for the trailer you want to change.

Except this last one, thanks Peter for the changes in the code of the simulator.

When I started this thread for early March I did not think it would
become one of the biggest threads here on the forum.
Friction on the train engages the people.
The last thing here that Peter Gulyas writes about sounds very interesting too.

Bye Jorgen

[steamer_ctn]

gpz, on 24 April 2014 - 09:51 PM, said:

Have you realized, that I also described a possible method to reach it above my sentence you quoted (through 2 posts), not just dropped in a vague idea? :oldstry:

Yes, thanks, I realized that you offered a possible solution.

However as suggested, I still believe that some additional research, thinking and investigation will be required.

For example, the following questions or thoughts need to be considered:

i) There are over 30 Davis equations quoted in the FCalc documentation - so these would need to be ultimately integrated into OR. Do they cover all the necessary combinations?

ii) As you have rightly suggested some means of distinguishing between the different vehicles types would be needed. - it would also be necessary to distinguish between different types of freight cars

iii) Is PassengerCapacity() parameter used universally in definition of railcars? If not, it could lead to misrepresentation of some railcars.

iv) As you have suggested some thought will need to be given to the variation of C factor in different circumstances.

These are just a few things that I feel need to be thought through. I suspect that there are others as well.

The current use of Davis factors allows players to address some of these issues themselves.

gpz, on 24 April 2014 - 09:51 PM, said:

The options ORTSDavis_A, ORTSDavis_B and ORTSDavis_C must be considered as absolutely temporary options, should never be advertised to public as a general solution. Moreover, handling of the "C" values of FCalc is also defective principally. (Which is not an unsolvable problem for a software in development, of course, just everyone must keep it in mind.)

The other aspect that needs to be considered is whether the quoted FCalc equations cover all the relevant possibilities.

For example the Canadian Resistance Paper quoted in this thread raises some interesting questions as well, as the Davis parameters are different yet again. Further down in the same paper, the AAR formula quoted also "deletes" the B value from the equation for certain track types. In addition the impact of different track types and articulated vehicles will also impact the resistance effect.

This page also quotes more variations to the Davis formulas, so do we exclude the possibility of including these values?

So whilst I acknowledge that your proposal addresses some of the aspects, I still believe that some further thinking and research is required to arrive at an appropriate level of consistency.

For the time being I believe that the ability to add the Davis parameters provides the ability for players to adopt some level of "customisation".

I have found the study of train resistance to be an interesting area, but it requires some detailed thinking and research, for example, I am currently trying to understand the resistance impact when a train enters a tunnel. This page describes the effect at a very high level.

Cheers

[steamer_ctn]

Hi Jorgen,

jorgen, on 25 April 2014 - 02:08 PM, said:

At last, the wagons now roll as they should. (as they did in Or 0.9)

Did some test. 51 Ton test wagons with these lines.
Friction (C1) 580N/m/s (E1)0 (V2)1.5mph (C2)2.61N/m/s (E2) 1.5

To make sure that we are on the same page, can you please post a snippet from your wagon file that has the friction definitions in it.

If you are NOT using the Davis values, and instead you are using the above "MSTS" values then the Bearing type, etc will NOT have any effect. See this thread.

If you are using the MSTS values, then nothing should have changed from OR 0.9. The bearing type is "embedded" in the FCalc calculations, refer to the FCalc documentation.

Thanks

[jorgen]

steamer_ctn, on 25 April 2014 - 10:47 PM, said:

Hi Jorgen,

To make sure that we are on the same page, can you please post a snippet from your wagon file that has the friction definitions in it.

If you are NOT using the Davis values, and instead you are using the above "MSTS" values then the Bearing type, etc will NOT have any effect. See this thread.

If you are using the MSTS values, then nothing should have changed from OR 0.9. The bearing type is "embedded" in the FCalc calculations, refer to the FCalc documentation.

Thanks

Hi The test i did. I use MSTS values
Friction 580N/m/s -0.10 1.7mph 4.201N/m/s 1 -1rad/s 0 1 )
But in test 2 to 4 i add this line ORTSBearingType ( Friction roller Low )

Today i did test with Davis values , have rewritten my post yesterday Link
The text from link
Here was the original Msts values
Friction 580N/m/s -0.10 1.7mph 4.201N/m/s
5.1N/rad/s 1 -1rad/s 0 1 )
ORTSBearingType (Low)
-----------------------------------------------------
Then I replaced it with Davis values
ORTSDavis_A (188.165)
ORTSDavis_B (2.49802)
ORTSDavis_C (0.045)
ORTSBearingType (Low)

The results up to 5 Mph was same, but the difference were then the train come up to 5 mph.
If using (MSTS values) the friction were 580 Newton
Using (Davis values standard freight) the friction were 188 Newton.
----------------------------------------------------------------------------

So you saying if you don't change in wag files
will function as the Or 0.9. Ie no start friction at all.
This is good ,but now its working for me. Se attach file.

When driving with this wag file x2194, you get 5000 Newton i start friction. (file 5,15 kb)
in or 0,9 is that 580 N (862 % higher friction)
But if you add the line ORTSBearingType (Low) on MSTS values (file 5,20 kb).
The start friction be 963 Newton.

Attached File(s)

•  3dts_sfwellcar_klinedouble.wag (5.2K)
  Number of downloads: 264
•  3dts_sfwellcar_klinedouble.wag (5.15K)
  Number of downloads: 249