German Train Physics Inaccuracies (freight Edition)

Discussion in 'TSW General Discussion' started by cwf.green, May 7, 2021.

  1. cwf.green

    cwf.green Well-Known Member

    Joined:
    Mar 4, 2019
    Messages:
    815
    Likes Received:
    1,346
    I already made a thread with my results of testing (primarily the brakes) on German passenger trains and locos in a previous thread in the Technical Reports section (found here) but since I will add these to a ticket and I wanted some input from players/drivers I thought to put it in the PC Discussion instead. Hope that is okay with the mods.

    TL:DR (I'll leave the smiley because it looks funny)
    1. Locomotives are modelled as using P-brake while wagons are in G-brake but this is opposite from reality. Brake cylinder pressure in locomotive should increase/decrease much more slowly.

    2. All wagons are in G-brake but only the five first wagons (on heavy trains) or none should be in G-brake, rest in P-brake with similar application and release rates as passenger wagons.

    3. All freight wagons except Laaers brake way way too hard. Some as much as 66% too much. I've provided all information necessary to fix this.

    4. Many freight wagons have incorrect weights, some are empty while in loaded configuration and some weigh twice as much as axle load would tolerate (180t per Falns wagon).

    5. Tread brake physics is incorrect, I've provided a table and description of how to fix this.


    In the thread above I detailed my methods from determining realistic braking distances/brake forces for trains so I will be more concise (spoiler: I failed) here (and I have some different feedback that may need to be explained instead).

    First off all the German freight trains in TSW2 behave as if they are in G-brake setting. The way it has been described to me by real world drivers is that the dumping rate and loading rates of the brake pipe is determined by length of the train, the valve area (larger area for emergency compared to service brake) and the compressor performance (when releasing the brakes i.e. loading the brake pipe). These rates will be constant no matter if the brakes are in G,P or R-setting.

    The G-brake setting instead determines how quickly the brake cylinders increase and decrease their pressure. G-brake has much longer application and release times (18-30s application and 45-60s release vs 3-5s and 15-20s for P and R).

    Anyone who has driven German freight trains in TSW knows that it will take much longer than 3-5s to reach full brake force and much longer than 15-20s to release a full service application, so we can conclude that the train is in G-brake.

    However, and this is the first inaccuracy: The brake cylinder increases and decreases in the locomotive as if the locomotive is in P or R-setting. This can be confirmed by running the BR185 or BR155 in single loco configuration. They act similarly to passenger locomotives. Instead they should apply and release in accordance with the timings I mentioned above. Right now they are basically doing the exact opposite of what the G-brake is for: to reduce couple forces in long and heavy trains.

    How would a realistic simulation look like in the cab? The brake pipe would react as it does currently but the brake cylinder gauge would lag behind.

    The second inaccuracy is that all the wagons are in G-brake setting (you can hear the last wagon release after about 55s) but this is very rare to occur. I know that the 9000 tonne "monsters" in northern Sweden (the iron ore trains) have all the wagons in G-brake and presumably the 4500+ tonne ore trains that run between Dillingen and the North sea ports also run fully in G-brake but much more commonly are the following settings:

    Trainmass: Brakesetting Loco /Wagons
    0-800t P/P
    800t-1200t G/P
    1200t-1600t Loco + first 5 waggons: G / Rest in P
    1600t-2500t Loco + first 5 waggons: G / Rest in P
    2500t-4000t Loco + first 5 waggons: G / Rest in P

    As you can see, all the trains in TSW2 would be G/P or "Langer Lok" i.e. only the first 5 wagons are in G. What this means for players is that realistic trains would be slightly more reactive and you could also see the slow release and application on the brake cylinder gauge in the cab.

    The third inaccuracy is similar to the passenger edition: All the wagons and locos brake too hard!
    Testing different freight consists I noticed that there is no difference in deceleration, they all have the same BrH or deceleration.

    From 100 km/h the BR185.2 stops in 410m which yields a BrH of 119 while the real one (in G) is 92 with a realistic stopping distance (no E-brake) of 520m. The brakes are 30% too strong.

    To test the wagons I took a heavy train (the 20 tank wagons) and applied full service (no E-brake) from about 120 km/h and then measured the stopping distance from 100 km/h (to make sure the brakes had fully applied when decelerating through 100 km/h) and then added the free run distance for P-brake of t/2 = 2s or 55m. Since the train was so heavy I could approximate the deceleration of an individual wagon as the deceleration of the train (the effect of the locomotive is negligible). The average deceleration rate of all wagons was 0.88m/s and the calculated stopping distance (with the added free run distance above) was 580m which gives a BrH = 81 for all freight wagons. Here is a table with (realistic) wagon masses, BrH and brake masses for the freight wagons used in MSB and RSN (I don't own RRO):

    Wagon Mass(E/L) Brake mass(E/L) BrH (E/L): Stopping distance(P-brake) (E/L)
    Falns 183 24.1t/89.6t 28t/58t 116/65 620m*/710m
    Shimmns-u 22.4t/89.4t 29t/59t 129/66 560m*/700m
    Shimmns-ttu 722 23.7t/89.7t 27t/51t 114/57 630m*/790m
    Roos-t 645 25.2t/80t 28t/54t 111/68 640m*/680m
    Habbiins 344 27t/90t 27t/58t 100/64 700m*/710m
    Laaers 560** 30.1t /54t 30.1t/54t 100/100 700m*/480m
    Zacns 22.6t /90t 26t /58t 115/64 620m*/710m

    E = Empty, L = Loaded
    (*): Stopping distances are from 120 km/h with free run distance included, otherwise from 100km/h.
    (**): Weights are for the 2 unit Laaers wagon.

    Average deceleration (application time excluded)(E/L) and relative errors:
    Falns 183 1.00/0.59 m/s^2 -12%/+50%
    Shimmns-u 1.12/0.60 m/s^2 -21%/+47%
    Shimmns-ttu 722 0.99/0.53 m/s^2 -11%/+66%
    Roos-t 645 0.96/0.62 m/s^2 -8% /+42%
    Habbiins 344 0.87/0.59 m/s^2 -1% /+49%
    Laaers 560 0.87/0.91 m/s^2 -1% /-3%
    Zacns 1.00/0.59 m/s^2 -12%/+49%

    Note: All stopping distances are for P-brake since this is what they would be tested at, in G-brake the stopping distances would be significantly increased but since most wagons would be in P anyway and since the deceleration rates are what DTG need to update, this is why I calculated them this way.

    As can be seen above, pretty much every wagon has ridiculously high brake force in comparison with real values (up to 66% wrong!) when loaded and slightly too low when empty, except for Laaers.

    At the moment TSW2 simulated freight trains as passenger trains in G-brake, basically completely unrealistic.

    The fourth inaccuracy is that many of the freight wagons have the same weights for loaded and empty configuration and many of them are wrong, sometimes incredibly wrong such as the Falns coal wagons which are listed as weighing 180 tonnes per wagon! The Habbiins wagons are always empty (weighing 27 tonnes). I know some of the listings are incorrect and the actual in game weights are more realistic but since I didn't want to test the accelerations of every wagon I feel satisfied with listing realistic weights above and DTG can cross check them.

    The fifth and final inaccuracy is something I have mentioned in a previous thread. Tread brakes with cast iron brake shoes (which most of the German freight trains use) don't produce constant brake force w.r.t speed but rather have increasing brake force as the speed decreases. This is due to the coefficient of friction (CoF) between the brake shoe and the wheel tread having a speed dependence.

    I found an article from 1937 (link is unfortunately in Swedish but the graph should be readable) which showed a graph of experimental data from a test of the CoF done by the German DR and using the average friction over the velocity span (which is approximately 0.09 for 0-100 km/h) together with the fact that

    F_brake = (brake mass)*g*(average CoF),

    gives a value for the average deceleration as

    a = BrH * g * (average CoF),

    where g = 9.81 m/s^2 or gravitational acceleration.

    Comparing these average deceleration values to the ones calculated above from the UIC formulas give almost perfect agreement! Perhaps this is not surprising since I was told that when the brake mass/BrH concept was first introduced the formulas and tables were produced from brake tests on German freight wagons and French passenger wagons back in the late 1920s.

    Below is a table of CoF vs speed (in km/h).

    V (km/h) CoF
    0 0.196
    10 0.139
    20 0.116
    30 0.102
    40 0.0915
    50 0.0832
    60 0.0767
    70 0.0721
    80 0.0684
    90 0.0656
    100 0.0629
    110+ 0.0610

    If DTG were to change the current unrealistic CoF dependence for the table below the only thing they would need to do to get correct deceleration would be to multiply the CoF value for a specific speed with the BrH * g.

    That is,

    a(v) = BrH * g * CoF(v)

    Or they can simply norm the curve above to be 1 at v = 0 (so available acceleration = max_acceleration * CoF(v)/CoF(0)) and infer the deceleration rate from the stopping distances I provided, although this is more tedious.

    I don't have high hopes that these fixes will be made, but I think that they could be made in a few hours by one person (If I've done a half decent job at explaining) and I truly believe these would greatly increase the immersion factor and enjoyment for players. Different freight trains would feel and behave slightly differently making gameplay more varied and increasing replayability.
     
    • Like Like x 12
    • Helpful Helpful x 3
  2. cwf.green

    cwf.green Well-Known Member

    Joined:
    Mar 4, 2019
    Messages:
    815
    Likes Received:
    1,346
    Can someone that owns RRO confirm whether the new freight wagons that are included are the Sdggmrss 6-axle twin-container wagons with Jacobs bogies? I have mass & brake mass figures for the Sdggmrss that I can add to the data above but I can't find any information about which new wagon was introduced with RRO.

    Also, if anyone has any questions about my methods or are interested in a clarification about something I'd gladly try to help.
     
    • Like Like x 2
  3. breblimator

    breblimator Guest

    Haha, You are crazy! Try some runs with non-dry track conditions. Good job \o/
     
  4. cwf.green

    cwf.green Well-Known Member

    Joined:
    Mar 4, 2019
    Messages:
    815
    Likes Received:
    1,346
    Haha :D I guess this is what I do when I'm bored lol. I also do it because I think the fixes aren't that difficult for DTG to implement (I'd say it takes significantly less work than implementing the PIS for example) but would make a significant difference in realism. Right now if you drive a freight train in TSW2 it doesn't really matter what wagons you have coupled since they all brake the same (difference in length will have an impact on brake application time but still small in comparison to my fixes). It would also mean that what PZB mode and what speed you keep would have a big impact in difficulty stopping in time for red signals/speed reductions or difficulty in preventing PZB penalty brake applications.

    About the different track conditions: I just tested a 1400 tonne tank train on HRR in winter + mist + full wetness and the deceleration was still around 0.87 m/s^2 which is too close to the value 0.88 m/s^2 I found earlier to conclude whether there is any difference. My presumption is that since the available traction in the most adverse conditions in TSW2 is still around 0.2 (so maximum deceleration is 0.2 * 9.81 ~ 2m/s^2), which is much larger than any wagon can manage, there will be no difference.

    For adverse traction conditions to affect the stopping distance of a heavy freight train the CoF has to drop below ~ 0.1.
     
    • Like Like x 3
  5. Inkar

    Inkar Well-Known Member

    Joined:
    Dec 13, 2016
    Messages:
    1,417
    Likes Received:
    2,335
    I think this is what you are asking for:

    pic1.jpg
    pic2.jpg
     
    • Like Like x 2
    • Helpful Helpful x 1
  6. cwf.green

    cwf.green Well-Known Member

    Joined:
    Mar 4, 2019
    Messages:
    815
    Likes Received:
    1,346
    Thank you. I couldn't find sggmrss in the table of brake weight's I've accumulated but I do have brake weights for the sggmrs wagon which is the 100 km/h variant of sggmrss (ss = 120 km/h max).

    The sggmrs uses a mechanism (GP-A) that scales the brake weight with the gross weight of the wagon up to some maximum brake weight (which is 102t for sggmrs). Since I could only see the sggmrss wagons in RRO loaded with two 40' or 45' containers (which have maximum gross weights of ~ 31t) the sggmrss probably won't exceed (29.5+31+31)t = 91.5t which both allows it to travel at 120 km/h according to the data in the second picture (although greenbrier website limits 120 km/h to 80t gross) and the automatic load-sensing means the BrH is 100 in both empty and loaded configurations.

    All this means that the stopping distance of a lone sggmrss unit (in P-brake and emergency) from 120 km/h will be 700m, giving it an average deceleration of 0.87 m/s^2 with tread brakes and a constant deceleration of the same value with disc brakes.
     
    • Like Like x 1
  7. breblimator

    breblimator Guest

    Perhaps I showed you this file before - I think it's all there. Enjoy your meal :)
    PS Yes, but there is no data on which ones are used in the game. Yeah, heh
     

    Attached Files:

    Last edited by a moderator: May 8, 2021
    • Helpful Helpful x 1
  8. cwf.green

    cwf.green Well-Known Member

    Joined:
    Mar 4, 2019
    Messages:
    815
    Likes Received:
    1,346
    Thank you!

    I have a notepad document with weights and brake weights but I found a link that pretty much has all my data already:
    https://zusiwiki.echoray.de/wiki/Sammlung_von_Bremsgewichten

    "Umstellgewicht" for those who don't know means the weight at which you should change from empty to laden or vice versa (otherwise you might find yourself with way too little brake force and then not even PZB will be able to help you, of course this should be noticed during the running brake test but there have been several accidents due to this error) on brakes that are manually set.
    "Autom. Lastabbremsung" is the automatic load sensing I mentioned earlier, the brakes will automatically match the brake weight to be equal to the gross weight so that BrH = 100, up to a maximum where they are fixed and you calculate BrH as with the manual variant.
     
    • Like Like x 1
  9. a.paice

    a.paice Well-Known Member

    Joined:
    Sep 10, 2020
    Messages:
    974
    Likes Received:
    1,589
    Regarding the RRO Sggmrs container wagons, I’ve found it takes about 5 minutes to release the brakes fully. I understand it will take a while but 5
    minutes seems far too long and it’s hard to slow down for PZB speed limit’s without overcooking the brakes and coming to a stand.

    I usually use the 30 wagon consist in the scenario planner- does anyone know if this is realistic?
     
    • Like Like x 1
  10. breblimator

    breblimator Guest

    Container / Intermodal / High Priority / Quality Train
    Well, now we know why. It has no brakes! :D

    EDIT
    ZUSI... well, it's still worth a look here
     
    Last edited by a moderator: May 8, 2021
  11. KoeleKoen

    KoeleKoen Well-Known Member

    Joined:
    Apr 28, 2020
    Messages:
    360
    Likes Received:
    684
    With the traxxes i turn off ep brake on back panel and and on sidepanel next to the desk i pick pn instead of el, to me it feels a bit more slower and realistic or is that just in my head?
     
  12. cwf.green

    cwf.green Well-Known Member

    Joined:
    Mar 4, 2019
    Messages:
    815
    Likes Received:
    1,346
    If you mean that it takes 5 minutes to release the brakes after you have made maybe a 1-1.5bar application then I feel very confident in saying that that is unrealistic, it should at most take around a minute (in G-brake).

    It shouldn't make a difference since EP brakes only work on passenger wagons. For passenger trains it should increase all timings on the last wagon by about 2-3 seconds. With el the signal travels at the speed of light while for pn the signal travels as slow as 100-150 m/s (not at the speed of sound due to flow resistance). Most tables I've seen add 1 sec per 100m.

    I did test el/pn on a 20 wagon tank train with the BR185 and there might actually be a difference in the game. I timed how long it take for the train to move after releasing the brakes from full service and it took almost exactly 1 minute (or 55-60s) with the el-setting but it took an extra 10 seconds with the pn setting. This is weird and wrong. Another bug I noticed was that the brake pipe gauge stopped working when the selector was in pn mode, it was stuck at 3.4 bar.
     
  13. KoeleKoen

    KoeleKoen Well-Known Member

    Joined:
    Apr 28, 2020
    Messages:
    360
    Likes Received:
    684
    Yeah i know, i drive with passengers irl so there it's always ep but it felt like the game also simulated ep on the freight, that's why i tested that, it felt like ep responsiveness
     
  14. cwf.green

    cwf.green Well-Known Member

    Joined:
    Mar 4, 2019
    Messages:
    815
    Likes Received:
    1,346
    I think you're right. It does look like the el/pn selector changes stuff on the TRAXX locos.
     
  15. tymon967

    tymon967 New Member

    Joined:
    Apr 4, 2020
    Messages:
    2
    Likes Received:
    2
    Fixing these inaccuracies would be a great step towards increasing the quality of simulation, however i think it will never happen because it will mess up all of existing timetables. All freight trains would have their runtimes changed.
     
  16. KoeleKoen

    KoeleKoen Well-Known Member

    Joined:
    Apr 28, 2020
    Messages:
    360
    Likes Received:
    684
    For the player you mean? Cause none of the ai use physics from simugraph, they also go full linespeed no matter what pzb mode they should be in.
     
    • Like Like x 4
  17. cwf.green

    cwf.green Well-Known Member

    Joined:
    Mar 4, 2019
    Messages:
    815
    Likes Received:
    1,346
    I don't think that is true (fortunately). Nearly all my fixes are about brake performance which plays quite a small part in terms of running time. Also, as KoeleKoen mentioned, I don't think the AI have the same fidelity of physics. If they did that would be terribly inefficient.
     

Share This Page