Friction settings

[MikeK]

A question for anyone who has played with these settings. Based on this info, it seems that the drag coefficient and rolling friction coefficient are too high on all the default rolling stock. Most have values of 4 and 0.0015. This makes sense to me since I notice strange things like having to drag trains down shallow hills.

I had a plan of using RW_Tools to change the values for all the US rolling stock and I have some questions for anyone who has dug in there before. Specifically, what are the interactions between these values that I found in one of my default locos:

Code: Select all

   <DryFriction d:type="sFloat32" d:precision="string">1</DryFriction>
   <WetFriction d:type="sFloat32" d:precision="string">0.5</WetFriction>
   <SnowFriction d:type="sFloat32" d:precision="string">0.3</SnowFriction>
   <SandFrictionMultiplier d:type="sFloat32" d:precision="string">2</SandFrictionMultiplier>

   <MaxForcePercentOfVehicleWeight d:type="sFloat32" d:precision="string">61</MaxForcePercentOfVehicleWeight>

What is a reasonable value for "MaxForcePercentOfVehicleWeight"? 61 seems way too high for me.

Looking at how the sim might be using these numbers, it seems that:

- on dry rails the wheels will slip when the force reaches Loco weight x 1 x 0.61 = 61% of the loco weight.
- on wet rails the wheels will slip when the force reaches Loco weight x 0.5 x 0.61 = 30.5% of the loco weight.
- on icy rails the wheels will slip when the force reaches Loco weight x 0.3 x 0.61 = 18% of the loco weight

But then is you add sand into the equation, doesn't this mean that:

- on dry rails + sand the wheels will slip when the force reaches Loco weight x 1 x 0.61 x 2 = 122% of the loco weight. <-- oops!
- on wet rails + sand the wheels will slip when the force reaches Loco weight x 0.5 x 0.61 x 2 = 61% of the loco weight.
- on icy rails + sand the wheels will slip when the force reaches Loco weight x 0.3 x 0.61 x 2 = 30.5% of the loco weight.

Obviously that can't happen in reality. Am I interpreting these values correctly? Does sand even make a difference on dry rails in reality?

What is a reasonable set of numbers for these 5 values for a locomotive?

What about for a freight car? (both locos and freight cars have these values in their files). I assume these apply to brakes on a freight car.

What does it mean to have a sand multiplier for a freight car? Is this number applied when the loco sanders are on, or is it ignored because freight cars don't have sanders?

[Kali]

A few things first:

* Units are all metric, force values will be in Newtons/areas in square metres. Other than tractive effort ( which doesn't actually matter unless you're working with steam engines, and then it's all that matters ) which is in kilo-lbf. Except when you're doing tractive effort in graphs, and then it's in kN.
* MaxForcePercentOfVehicleWeight is for brakeforce only.

RW adhesion is thoroughly broken, but those figures are way too high anyway; the friction values are between 0 and 1 - or at least the blueprint editor won't go higher than 1 - so 1 there is presumably 100%. RW seems to take the adhesion weight of a locomotive and do something like divide it by the number of axles in the entire consist at scenario start, so adhesion varies with consist length... so you have to make it unrealistically high or start every scenario light engine. 0.5 for dry rails is probably a sane base figure there for a locomotive which is expected to pull a US-sized train, even if it's way too high for real life.

Now for the other part, brakeforce; you can get these figures fairly easily. It's complicated by "vehicle weight" always being the entry in the <Mass> section and not counting the vehicle load - really every reasonably recent brake system auto-adjusts for car mass - but for a US freight car it's 10-14% ( and 14% is for something fresh out of the shops ). For european freight stock it'd be more like 55%, european passenger stock is more like 65-75 depending on brake systems so I would imagine recent US passenger stock is similar. Locomotives depend on the locomotive, but that data is usually published somewhere. Don't forget if you're setting a freightcar up that you have to decide if it's loaded or not; it's probably best to assume it is, so then your brakeforce entry needs to be multiplied by the proportional increase in mass so it works when the car is loaded. It's way too effective empty then, but either you provide empty & loaded cars or you have to pick one.

Worked example:

Target brakeforce: 10%
Empty mass: 25t
Cargo mass: 100t

Brakeforce entry: 0.1*(125/25) = 50% <-- this is what you'd put in the blueprint
Crosscheck: 25t*0.5 = 12.5; 125/12.5 = 10% <-- the resulting effective brakeforce when loaded.

[MikeK]

Thanks Kali!

It sounds like the only way to get realistic braking forces is to have separate empty and loaded blueprints. I hope that this is something RSC can address soon. Since wheelslip is clearly not modelled properly, I think I am going to use values that are high enough that I won't see any wheelslip at all.

I started looking at what was in the simulation file for locos to see how the brakes are set up. After reading this page, it seems that using only loco brakes should be enough to hold a 4 loco/15,000 ton train on a 1% grade. In RW however, the loco brakes are worse than useless. I figured this is a nice round set of numbers to use for testing.

However, what I found in the simulation file has me totally confused. Look at these values in the SD70ACe bin file:

MaxPower = 4300 // Makes sense, it is a 4300 HP loco
MaxForce = 191 // After a bit of searching I found a starting TE value of 191,000 lbs for the SD70, so I guess that is what this is
MaxPowerAtRails = 4020 // I suppose this accounts for drivetrain losses

But then in the handbrake section:

MaxForce = 15000 // What? The brakes are 78 times stronger than the motors?

It gets even stranger when I looked at the ES44 sim file:

MaxPower = 3100 // Shouldn't this be 4400, since the SD70ACe had 4300?
MaxForce = 198 // 198,000 lbs?
MaxPowerAtRails = 3100 // No drivetrain losses?
Handbrake->MaxForce = 15000 // And yet they don't hold the train in the sim.


What is going on here? Why does the SD70ACe have a MaxPower value of 4300, but the ES44 is only 3100?

When I look at the TractiveEffortVsSpeed (or Throttle) curves for both locos, they peak at 880. Where did this number come from?

[Machinist]

All of this was already deeply discussed earlier, I guess mainly in the "Dodgy Physics" thread, but in many others also.

According to Engine's Information Data (in game) ES44AC has 4,400HP (british, or 3,284KW). However, the blueprint are in KW, not in HP instead, so when you have 4400 for the "Power" blueprint value, in fact the engine is set to 4,400KW or 5,900~6,000HP (it's overpowered in fact, not the contrary).

ES44AC value I would say is quite correct once 3,100 (is KW) is equivalent to 4,160~4,200HP, which is very close to the expected HP of a brand new ES44AC (not all power is transmitted though, you can discount about 8%, engine's mileage dependent). The problem is the whole dodgy physics of game, including power, force, power at the rails, traction, tractive effort, friction's boggies, mass, weight, drag coefficient etc. etc. etc. The list is huge and all those paramters should be perfectly synchronized and tweaked together, just tweak one (or some) of them is for no avail.

Bottom line: at least for the game's parameters, ES44AC is likely correct (3100KW ~= 4200HP, I'm rounding once there are different HP measures) and SD70ACe is overpowered (4,400KW ~= 6,000HP). ES44AC is not messed; either SD70ACe and the whole physics are messed in fact.

You might look into the RailWorks user's wiki (missing a lot of data) but still usefull.
http://www.railsimdownloads.com/wiki/tiki-index.php

Doc.

[Kali]

I've started using code that outputs a known TE figure at a given speed, which you can get off a TE graph ( obviously depending on RPM/electrical system state )... that way it doesn't matter what units the power figure is in as long as the number is big enough. Having said that, 1kW is 0.75 of 1bhp, so for a rather worn out SD40/GP9 or something that's actually not a bad figure even if it's in BHP, accounting for transmission loss & engine wear. You can roll your own TE graph from an existing one, for a constant power machine they're all the same shape and the real max continuous TE figure is usually where the straight line bit meets the curve, you just have to adjust the X/Y units to match the gearing once you've positioned the curve corner in the right place. I have test results I've done again which come out with MaxPower being in BHP or kW depending on test, I don't care anymore :P

MaxPowerAtRail appears to do nothing at all for any blueprint unless it's AI. MaxForce on a diesel electric is seemingly only there to calibrate the ammeter - that is the TE value which will make the ammeter reading max out. I think it might be something the AI reads too but I've not got round to investigating the AI. It's also the only blueprint figure a steam engine seems to depend on...

Handbrakes are in lbf according to the BPE, and are also seemingly totally independent of the brakeforce the associated brake can produce, MaxForce is in k-lbf ( the force figure in the dynamic brake section is also in k-lbf ). The figures in TEvsThrottle are capping figures, they just have to be big enough that they don't cap output to something too low. The entire D-E model is horrible and I wish it wasn't far too late to scrap it and use a modified electric one, but that would mean redoing every existing diesel :(

[MikeK]

So if I am reading this correctly, the config files have both metric and imperial units mixed in together?

[Machinist]

So if I am reading this correctly, the config files have both metric and imperial units mixed in together?

That's correct, in the wiki you have the measure pattern used in each case. I guess that whatever relies on continuos and intense calculation is metric, and values for information/reference (or even simple calcs) may be imperial, not sure though.