This is true, but! 1k tons 500 m train vs 1k tons 750 m train? Is this just "weight-based resistance" or just more detailed?
Generally it seems a 3rd rail EMU slows down faster on the level then an over head EMU or a DMU, so I always assumed there was resistance based on number of contacts with the rails (IE a 3rd has extra contact points due to the shoes). Thats based on nothing concrete just my personal experience and view
I actually never had the feeling that trains are slowing down more when in curves. I've watched a stream from 'DadRail' where he drove a 66 (a Loco he drives in real life) there was one place (believe it was somewhere on SEHS) where he had to break on a slight downgrade in a curve. He told that irl he actually has to power up because those curves are pretty tight and slowing down the train a lot. So maybe there is a bit of that simulated but i dont believe its completely realistic. Its an interesting Thread though, maybe a Real Life Train driver will notice and tell us if its matching reality.
It’s definitely not enough and doesn't seem to vary with rolling stock. It makes sense that a heavy 66 with a train of battered old wagons should encounter more resistance than a light javelin for example. IRL some corners will have flange-greasers to ease the wear on the track and wheels which reduces the resistance where fitted.
In real life, AC traction motors create more resistance than DC traction motors. The confusing thing is that a unit running on AC power can have DC traction motors and vice versa, the power being converted before being fed into the traction motors. Generally speaking older EMUs - such as the 314 - had DC traction motors, even if the unit ran on 25Kv AC from the OLE. Newer EMUs - such as Electrostars - have AC traction motors, even if the unit is running on 750v DC from the third rail. I should point out that I’m not an engineer, so please feel free to jump in and correct any of the above. However, what I can say with certainty is that the difference when coasting is noticeable. A unit with AC traction motors loses speed noticeably more quickly on the same section of line that a unit with DC traction motors.
Worst example of poor inertia was the Class 70 in TSC. Running light engine, shut off power and it was like running through treacle it stopped so quickly. When I reported it go DTG my observations were dismissed and basically told it was a fudge so the physics worked when actually hauling a train.
I was thinking about this on the updated SEHS. I know in some places that the grades are quite steep but after driving for a while it felt weird that the train free wheels gaining speed on almost every downhill section, even when doing 225 kmh. The air resistance at this speed should be huge. It will take someone with first hand driving experience to clarify if it's accurate or not.
Considering some of the gradients on SEHS are as much as 1:40 I can easily see a javelin being able to have the speed increase without power though in real life I have never been on a line which permits more than 125 mph so that 15 mph could make a difference. Overall I think there is definitely rolling resistance in game and does a decent job but it just isn't as sophisticated as what it is in real life.
Simugraph (and in turn TSW) simulates rolling resistance, and air resistance. Since rolling resistance as a force is proportional to the mass of the train, the deceleration that rolling resistance contributes will not depend on weight. If you have two trains with the same type of vehicles but one of the trains weighs 500t and the other weighs 1000t you will need twice as much tractive effort to keep the heavier train from decelerating, but with no effort applied they will decelerate at the same rate and stop in the same time or distance.
Hi, I had the opportunity to drive some DMUs, and I felt a little more deceleration than in TSW, especially in curves. Tighter curves indeed have more impact on the deceleration. Keeping the throttle on the first notches was indeed required.
Curve drag (or curve resistance) is not currently (to my knowledge and I'm pretty confident about this assertion) simulated in TSW, but there is no reason in principle for why it couldn't be implemented in the future. It's quite an interesting physics problem. From my understanding the wheel flanges contact the sides of the railhead in curves. The wheel flange is pressed against the side of the railhead with a force that is increased with higher speed and/or smaller curve radius (for the same flange geometry). It also depends on superelevation. Depending on the contact area you have a certain coefficient of friction so there is a friction force that decelerates the train. Apparently (according to Wikipedia) several studies were performed in Russia on this problem and the resistance seems to mostly depend on the radius (inverse dependence) of the curve rather than speed. Interestingly, the curve resistance is actually minimized at some specific speed for a given radius and then increases for speeds above and below this speed. This is surprising to me since I assumed that centrifugal acceleration would be the main cause of curve resistance but then I remembered that curves are usually canted (superelevation) so for low speeds the wheels will ride the inner rail (causing resistance) and for high speeds they will start to ride the outer rails and at some speed inbetween the forces (gravitational and centrifugal) will balance and you basically have no push to either side, minimizing the curve resistance. My guess is that for no superelevation the force is minimized at low speed and then increases monotonically (math term for a function that never goes from increasing to decreasing or vice versa). Another hugely important type of resistance (which is simulated in TSW) is air resistance. At higher speeds the total resistance is dominated by air resistance. So for high speed trains it is of great benefit to try to make the train as aerodynamic as possible, even at the cost of some weight (which would increase the rolling resistance energy losses). Note: Some readers may object to my use of the term "centrifugal acceleration/force" and point out that centrifugal forces are fictitious. To them I have this XKCD comic.
However, it is certainly true that in, say, shunting maneuvers, coasting on the level will definitely bleed off speed, in cases where air resistance would be close to nil. Now, that may be rolling resistance built into the axle bearings, but something is producing drag.
Usually the model used for train resistance is something called the Davis equation (or a similar variant). In this model the train resistance (typically in Newtons per tonne or lbf per U.S. ton) can be written as R = A + B*v + C*v^2. Here v is the velocity (in some units) and the A-term corresponds to rolling resistance, which as you said is caused by the friction in bearings and also deformation of the rails. This term can also include grade drag: I*g, Where I is the percentage of the gradient multiplied by some constant to get the units to match and g is gravitational acceleration. The A-term can also include curve drag on the form K/R + L where K and L are constants and R is the curve radius. Typically (from what I have found), these latter two contributions aren’t included in the A-term of the Davis equation since the equation is supposed to describe resistance in flat and straight travel and those are instead added separately. Then we have the C-term which corresponds to air resistance and can also be written (in SI units) as 1/2 * rho * C * A * v^2 / M which might be familiar to those with knowledge on fluid mechanics. This is simply the drag (normalized by the mass M) from turbulent flow. Rho is the air density, C is here the drag coefficient (not the same C as in the Davis equation) and A is the surface area. So far all of these terms are used in Simugraph, although they are calculated as forces, rather than resistance (unit of acceleration) and the summing of the drag terms is per vehicle in the rake but with different drag coefficients from “coupled” and “uncoupled” areas (i.e the front has more drag than the ends of two coupled wagons). The last term is, to my knowledge, not included in Simugraph. This term is quite interesting since it doesn’t have a solid physical explanation. In fluid mechanics such a linear term would describe laminar flow, but this should be a negligible contribution in reality (laminar flow is more or a thing in slow viscous movement, think dropping a steel ball in syrup). My guess is that it is supposed to describe some form of speed dependence in the friction of axle bearings. I have also seen research where this term was included in the model and curve fitting of the train resistance gave a negative B-constant. This seems unphysical. Nonetheless it is often included in US equations and excluded in European ones. Discarding it is usually not a big deal since you can modify the other terms to get a resistance function that looks very similar to one with the B-term included.
