Correction To Earlier Findings Of Brake Physics

Discussion in 'PC Discussion' started by cwf.green, May 18, 2021.

  1. cwf.green

    cwf.green Well-Known Member

    Joined:
    Mar 4, 2019
    Messages:
    476
    Likes Received:
    670
    From my previous threads where I tested the physics of German freight trains I presumed that all the German freight wagons brake with the same deceleration. I made this assumption after testing 3 different consists which all braked the same. Upon further testing I have found that this is not true for all consists.

    TLDR: Of 5 tested freight consists/loco, 4 of them have significant errors in the brake physics. I outline how these can be fixed with just a few parameter changes.

    Testing methods:
    I found a section of at least 1km leading up to a signal, with straight track and zero gradient. A couple of hundred meters before this 1km section I braked the train from 130 km/h in full service, without electric brakes, and recorded the distance to the signal when passing 100 km/h and when stopped. The reason for braking from 130 km/h was to have the brakes fully applied at 100 km/h so that a very precise and accurate braking distance can be determined. Adding a free run distance to this measured distance I then calculated average deceleration and compared it to the real world values.

    Since the TSW2 BR185.2 has incredibly strong brakes (even stronger than the real locomotive has in R-brake setting) I had to account for this added brake force when calculating BrH for different consists to be able to compare the wagons without influence from the locomotive.

    NOTE: I have corrected for the deceleration caused by the locomotive to calculate the actual deceleration of the wagons, this is why except for the first entry in the table below all the consists are for the wagons only. As such the mass value is for the wagons only. I also decided to leave out brake mass and BrH because these are not necessary for DTG to fix the brake physics, instead I only have actual and expected average deceleration and stopping distances. I wish I could just uncouple the wagons and measure the stopping distance (or "bail off" the loco brake) but the train derails lol.

    Here are my findings for a few consists (L = load, E = empty):

    Consist------------------Mass-----------Average deceleration----------Stopping distance--------Error------------------
    BR185.2 (P-brake)---84t--------------0.87/1.43*-------------------------500m/325m*---------------66% too strong----
    10 Mixed Freight-----895.6t----------0.60/0.70*-------------------------700m/610m*---------------17% too strong----
    15 Habbiins (L)-------656t------------0.91/0.85*-------------------------480m/510m*----------------6% too weak-------
    10 Shimmns-u-------894.1t----------0.60/0.40*-------------------------695m/1020m*--------------32% too weak------
    15 Zacns--------------1350t-----------0.59/0.79*-------------------------710m/540m*----------------34% too strong----
    11 Laaers--------------592.3t----------0.91/0.72*-------------------------480m/595m*----------------20% too weak----

    (*) actual sim values.
    I will update with more consists when I get time to test them.

    Note: At least the Habbiins consist has an incorrect weight listed in the description (540t) or the locomotive tractive effort is too low since the acceleration would imply a weight of 740t. Perhaps many more consists weigh differently compared to their descriptions. In this case some of the values above will be incorrect proportional to the error of the listed weight. I'm trying to wrap my head around how DTG could make such a simple mistake.

    Conclusion: The good news is that DTG actually seem to try to implement different brake physics for different types of trains. The bad news is that the physics is very incorrect for all but one of the tested trains.

    My hope is that DTG will implement better brake friction curves as described in the freight thread. From there they would simply need to adjust all the wagons so that their maximum deceleration (w.r.t speed) would be

    a(v) = µ(v) * (Brake mass / Wagon mass) * 9.81,

    where µ(v) is the coefficient of friction as a function of velocity (in km/h) as described in the link above.

    As a slightly quicker (and hopefully temporary) fix they could simply increase or decrease brake force by the percentages I gave in the table above. This would give correct stopping distance from 100 km/h (but not from other speeds since the brake curve is incorrect).

    Lastly, they really need to change the application and release rates to P-brake on the wagons (and G on the loco). Right now it is the opposite which is completely unrealistic and also quite annoying to drive with.


    If DTG do not feel that these fixes are of high priority or cumbersome to implement, I volunteer to fix them myself for free (I'll sign the NDA, no problem lol). What do you say DTG Protagonist ? ;)

    PS: For those that question my sanity in doing all these tests, all I can say is that we all have different ways to procrastinate from what we should be doing :D

    EDIT: Added Laaers and fixed some typos.
     
    Last edited: May 20, 2021
    • Like Like x 13
    • Helpful Helpful x 1
  2. breblimator

    breblimator Well-Known Member

    Joined:
    May 1, 2019
    Messages:
    3,433
    Likes Received:
    3,977
    I have no words /o\ I am also signing your petition!

    You are probably among the doubters, but when you finally buy HHL: 143 vs 112, please :D
     
    • Like Like x 2
  3. cwf.green

    cwf.green Well-Known Member

    Joined:
    Mar 4, 2019
    Messages:
    476
    Likes Received:
    670
    Thanks breblimator !

    I've updated the table above with my findings of the Laaers wagons (the BR185.2 + 11 Laaers consist to be exact). The wagons brake about 20% too weak compared to the real ones. These wagons actually had the correct mass in the description.
     
    • Like Like x 1
  4. cwf.green

    cwf.green Well-Known Member

    Joined:
    Mar 4, 2019
    Messages:
    476
    Likes Received:
    670
    I just tested the x20 Shimmns Vans A and B in loaded and empty configurations. The listed train weights imply that both empty and loaded trains have the same weight (23.6t and 22.4t respectively) and that they are always empty.

    However, when driving the trains their accelerations imply that they all weigh around 900-1000t or about 43t per wagon. Because of this I don't feel I can accurately test how much they differ from the expected brake force (do I compare them to loaded or empty wagons etc).

    I'm also curious of how brake force is modelled in Simugraph. Do each vehicle get assigned a maximum brake force or a maximum deceleration (as in TS2021) that will then be reduced depending on what fraction of maximum brake cylinder pressure is applied?
    Are the brake forces per axle rather than vehicle?

    Perhaps TrainSim-Matt would like to answer this question? :)
     
  5. cwf.green

    cwf.green Well-Known Member

    Joined:
    Mar 4, 2019
    Messages:
    476
    Likes Received:
    670
    I hadn't noticed this before but the brake cut-in button is modeled in the BR185.2 which means I can accurately test the brake force of the wagons without the influence of the locomotive brake. I will test the brakes in this fashion and compare the results with the values I got above, I don't think they will change much, however, since I already corrected for the loco brakes..
     
    • Like Like x 1
  6. breblimator

    breblimator Well-Known Member

    Joined:
    May 1, 2019
    Messages:
    3,433
    Likes Received:
    3,977
    You've done so much interesting research already, I think it's time to produce some online sheet with results :) just saying \o/
     
  7. cwf.green

    cwf.green Well-Known Member

    Joined:
    Mar 4, 2019
    Messages:
    476
    Likes Received:
    670
    Thanks! I agree about the online sheet. I'll probably make a google spreadsheet and link it.

    I have bad news about the driver brake cut-out button by the way. It disables the brakes on the whole train so it is kind of useless for my purposes.
     
    • Like Like x 1
  8. TrainSim-Matt

    TrainSim-Matt Executive Producer Staff Member

    Joined:
    Nov 17, 2016
    Messages:
    1,081
    Likes Received:
    5,932
    Can you do your tests with the trains/stock that come on Hamburg as those are the latest and received updates along the lines you're discussing - and see if they are performing more to your expectations?

    Thanks
     
    • Helpful Helpful x 2
    • Like Like x 1
  9. cwf.green

    cwf.green Well-Known Member

    Joined:
    Mar 4, 2019
    Messages:
    476
    Likes Received:
    670
    Hi Matt. Thanks for responding! I will report my findings for Hamburg as soon as I'm able to test it.
     
    • Like Like x 2

Share This Page