Factorio Forums

When using the diesel locomotive, I couldn't but to notice a strange behavior:
Currently, driving close to full speed results in the least fuel usage while driving at low speeds actually costs more fuel overall when driving the same distance.

Expected behavior would be having the lowest fuel usage when driving (almost) at the lowest speed, since rolling friction is a constant, but drag should scale in square with the current speed. Regular friction can be mostly ignored since it is superseded by drag at higher speeds and by rolling friction at lower speeds.

Having drag as a square of speed would result in the expected behavior: The sweet spot for fuel usage would be at the equilibrium between rolling friction and drag, while high speed trains would consume far more fuel.

Current fuel consumption would be perfectly reasonable for operating a train at it's sweet spot, but running a train at full speed (that means only 1 locomotive) should balance out at about 4x increased fuel usage.

As a side effect a slider to adjust the target speed for the automated train network would be necessary.

Ah. I think the first is, because when the train is heavy (many wagons), it cannot reach the top speed (drives slower) and motor is running with full capacity. But if it is lighter, the train reaches the max-speed and releases the power-toggle a bit. For me this is logical. I can see that also on the number of smokes coming out of the loco.

The second is a good idea.

All in all I think the trains do use too less fuel. I made a 6000 tiles track and it is really enough to fill the trains on one side. I mean they drove 12000 tiles, take about 5 minutes for that and come back and just need some coal.

Actually I had to double-check (hi tt-forums) and apparently yes, this is sound logic. The faster you go, the less the environment you're in wants you to accelerate. If you let go off the pedal, you lose speed faster at higher speeds due to drag.

It's a perfectly common mistake in such simulations, developers only think about rolling friction and regular friction, but completely forget about drag.

There's also a sweet spot for efficiency in such system, but it's at full throttle. That's since the energy lost to regular friction is constant over a given distance, unrelated to the speed and the energy lost to rolling friction is constant in time and therefor inverse proportional to speed.

Btw.: OpenTTD has no drag simulated either, but it doesn't matter for them because they are not dealing any respect to fuel since it's dwarfed by other operational costs. KSP on the other hand gives you a lot of opportunities to experiment with the effects of drag.

Please excuse the off-topic:

ssilk wrote:All in all I think the trains do use too less fuel. I made a 6000 tiles track and it is really enough to fill the trains on one side. I mean they drove 12000 tiles, take about 5 minutes for that and come back and just need some coal.

I would really like to see some screenies of that factory and the map Had that long route some practical use? I was just thinking about suggesting "spreading" the resources and monster spawns more so you need to go out and build a train network at some point. You're pretty much set up locally at your spawn point as far as i can say; until rocket defense, there isn't too much need to build far away outposts, or am i wrong?

Well this are some weeks ago: https://forums.factorio.com/forum/vie ... ast#p21595

The latest saves are here: https://forums.factorio.com/forum/vie ... f=7&t=3524

The sense: cause I can.
No, the reason was, how it feels and how the gameplay behaves, when going into that big dimensions. Some of the suggestions are coming from there, for example how we can supply such far away areas. What's needed to find ores. How could the whole process be automated. Such things.

The spreading is exactly, what I mean would bring the game into a direction, which brings real flow into it, after you built the rocket defense, but only when it is much more automated! Because the need of time becomes more and more a problem then. So I mean, what's in the beginning is a very puzzle like game on small space changes then more and more to big areas of "nothing" and some points of "everything". Space is no more problem, the problem is the repetition.

Because after you built the rocket defense you need - in my opinion - to bring down the settlers. You need rockets, immense amounts of housing (just a number : what I think about are houses which can contain up to 1,000 settlers, and there are 1,000,000 on board), defrosting engines, medical labs for the settlers, gigantic amounts of food... The target is logically then to bring all down and to have survived as many as possible.

Again, nothing of that is somehow official, but I feel the direction goes there.

Weight doesn't reduce air friction and doesn't increase max speed (except its rolling downhill)
Engine power only does increase max speed

Ext3h wrote:Expected behavior would be having the lowest fuel usage when driving (almost) at the lowest speed, since rolling friction is a constant, but drag should scale in square with the current speed. Regular friction can be mostly ignored since it is superseded by drag at higher speeds and by rolling friction at lower speeds.

Ext3h wrote:It's a perfectly common mistake in such simulations, developers only think about rolling friction and regular friction, but completely forget about drag.

First of all, and wayyyyy more important than anything else, this is a game. It can use real life as a source of inspiration but is under no obligations to accurately match the real world.

Second of all, and the reason for the giant facepalm, is because if you were to use real life as a reference you'd find the current implementation is pretty accurate.

So, to run you through some of the fundamental physics here:

By their very nature, trains have extremely low energy loss to friction (both drag and rolling friction). Talking about them as a dominant force in energy consumption shows that you don't know what you're on about. Steel wheels on steel tracks don't have much cause for deflection (compared to saggy-ass car tyres) so rolling friction is tiny. Drag is also tiny relative to the mass of a train because all the carriages are tucked nicely behind the engine, so only the surface area of the front of the train has to push through air. Sure, air along side the train causes drag but that's even tinier because.. well aerodynamics.

Let me explain where the energy goes: F=ma. Force = mass * acceleration. Trains have a giant-ass mass. When a stationary train starts to move, 100% of the energy is dumped in to accelerating the enormous mass of the loco and any carriages.

Now, when it's at top speed you've stopped accelerating. So you don't need to consume any energy except enough to fight off internal friction, drag and changes in direction (vertical and horizontal).

Ok, so we've established two points on the energy consumption graph: At velocity=0, energy consumption is at maximum (engine gives everything it can) and at velocity=max, energy consumption is at minimum (engine only has to work against friction, no acceleration work required).

Now for the in-betweeny stuff. As the train gets up to full speed, so too does a component of the energy the engine generates get consumed by the opposing forces of drag. So the amount of energy left over for acceleration becomes a smaller and smaller. But even just before max speed, the component lost to drag is waayyyy tiny.

And finally, if you want to get really technical, the engine won't convert fuel in to mechanical motion uniformly across all velocities. It varies according to the power efficiency profile of the engine. Different engines have different power efficiency curves but the engine in game claims to be a diesel unit (that can somehow explode wood ). Generally engines are optimised to hit the peak of their power curve wherever the engine will spend most of its time doing work. That is: cruising speed. Some engines have a broader profile so that they can deliver good efficiency at a range of different speeds. This is the main reason for having a gearbox. Anyway, relating this back to the game it appears as if the engine is speed limited and it's easy to imagine that at this maximum speed the engine is operating at the peak of it's power curve (using whatever gearing is required).

Khyron wrote:First of all, and wayyyyy more important than anything else, this is a game. It can use real life as a source of inspiration but is under no obligations to accurately match the real world.

This is right, but there is no reason for not copy the drag behavior. Ok, there are much more important this. But: The unrealistic behavior is not part of gameplaybalancing. It wouldnt change much. Only be more intuitiv.

MisterSpock wrote:This is right, but there is no reason for not copy the drag behavior. Ok, there are much more important this. But: The unrealistic behavior is not part of gameplaybalancing. It wouldnt change much. Only be more intuitiv.

i think i can see where you're coming from but i disagree, the trains operate in a perfectly logical way in every aspect needed for the game
Accelerate -> train uses heaps of fuel to increase its velocity, the heavier it is the more energy it takes, which makes sense and obeys the laws of physics (i.e. momentum)
Stop Accelerating, -> train gradually slows down until a full stop is reached due to friction.
Reach Max Speed -> after reaching maximum speed the only act of the engine is to combat friction, it stops using energy to increase its velocity, so it uses the least amount of fuel

as far as my knowledge goes these trains work pretty much the same as how real ones work, any inconsistencies are probably due to restrictions in the game engine.

Khyron wrote:the engine in game claims to be a diesel unit (that can somehow explode wood )

a diesel engine which doesn't use petroleum or any oil products/liquid fuel. yea.... sounds more like a steam engine really

Dakkanor, great idea.

Petroleum should be used as a fuel too with cars and trains fueled by petroleum barrels or maybe a pipe to the train station and solid fuels used for smelters, burners and boilers.

Is petroleum a liquid or a gas in the game btw? It's called "petroleum gas" but you fill *barrels* with it.

Time for facts:
http://www.inrets.fr/ur/lte/publi-autre ... b_rail.pdf

Air resistance actually DOES exceed mechanical resistance for real life trains, especially once speed passes 100km/h or more. Even with perfectly shielded trains like the German ICE or French TGV postal, speeds higher than 200km/h are uneconomical, even though higher speeds would actually be possible.

For unshielded good wagons (or even worse: hopper wagons), break even point is actually as low as mere 50-100km/h, anything above causes a significant amount of air friction. There is a good reason why unshielded trains usually don't drive any faster than 120km/h top, even when both tracks (including increased distance between signals to account for increased breaking distance) and wagons could handle faster speeds.

How do you know that current train top speed is locomotive top speed? Usually train top speed is limited by track speed limit and it is far under locomotive top speed. So they don't need to use much throttle to keep speed. I think more unrealistic is that train isn't slowing down before curves.

I've never even noticed any of this. Automate the coal input of the wagons with Requester Chest at every train stop (only in my base) and all trains runs at full coal supply.

I begun to have only one coal supply per world. Train repeatingly drives it's targets, and after the fifth repetition or so, it stops to refill coal. Should be an option to say "repeat this route until ..., then switch to that route (which is in that case go to coalstop and switch back to first route)"