News:

Use the "Forum Search"
It may help you to find anything in the forum ;).

More realistic weight limits?

Started by MCollett, May 29, 2012, 01:49:43 AM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

MCollett

Weight limits, specified in the pakset by max_weight, are currently applied per vehicle both to ways and to bridges etc.  This is wrong for both.  The weight limit for a bridge should be based on the weight of the entire convoy (or perhaps the weight of one tile's length of the convoy), while the weight limit for a way should be based on the weight per axle

As a simple example of the inconsistencies that result: a 2-6-0 locomotive with 4-wheel tender can currently run on rails and over bridges that are forbidden to a 2-6-4T locomotive of the same total weight; in reality the available routes would be exactly the same for both.

I'm fairly sure that it would be straightforward to code this change, but obviously the weights specified in the pakset would have to be adjusted accordingly (up for bridges, down for ways).

Best wishes,
Matthew

sdog

this was discussed thoroughly a while ago, when wheight limits were introduced. For practical reasons the weight of the heaviest unit is used right now as an approximation.

Since this is on the todo list for standard, refinements are afaik not planned right now. As a good compromise some system with weight classes (eg. route availability) was favourably discussed. search engines should bring up the old threads.


ps.: welcome to the forum Matthew

jamespetts

SDog is correct - the issue of weight limits is not a straightforward one. This was discussed before, and the consensus was, if I recall correctly, that it would be too difficult for most users to understand if there were three sorts of weights about which the users needed to think: vehicle weight (for determining how many vehicles to add to a train/convoy before it is too heavy to pull by the chosen locomotive); axle weight (for ways) and convoy weight (for bridges), the latter of which would be especially complicated, since only that part of the convoy that is actually on the bridge at any given moment would need to be weighed - a bridge of four or five meters long over a stream or even of a dozen or so meters long over a road does not have to take the entire weight of a long loaded freight train.

And also welcome to the forums! It is always worthwhile seeing new people here.
Download Simutrans-Extended.

Want to help with development? See here for things to do for coding, and here for information on how to make graphics/objects.

Follow Simutrans-Extended on Facebook.

MCollett

Thanks to both of you for the welcome.

Quote from: jamespetts on May 29, 2012, 10:47:34 AM
the consensus was, if I recall correctly, that it would be too difficult for most users to understand if there were three sorts of weights about which the users needed to think
...
only that part of the convoy that is actually on the bridge at any given moment would need to be weighed

The latter point was why I made the alternative suggestion of 'the weight of one tile's length of the convoy'.  Going with this approach, we could also simplify by making the approximations of

  • calculating weight-per-axle assuming a standard axle-spacing, expressed as axles per tile, and
  • using the weight-per-tile for each vehicle rather than for the convoy as a whole.
Then the three weights would reduce to two: the limits for both ways and bridges could be expressed in terms of the vehicle weight per tile, i.e. the gross weight divided by the vehicle length.    Even this simplified version is still much more accurate than the current system: it would for example fix the tender vs. tank discrepancy I noted, since the extra length of the tank loco compared to the tender one would be roughly in proportion to its extra weight.

Adhesive weight is of course a whole different can of worms ...  .  (No, I'm not serious: to the extent that adhesion is relevant it is already adequately subsumed in game terms by the tractive force.)

Best wishes,
Matthew

ӔO

Personally, I thought weight class was a better system, but only because it would have been fairly easy to use. As a downside, the dat files might be a bit harder to create.

class 0: can go everywhere
~
class 10: only class 10 or higher
My Sketchup open project sources
various projects rolled up: http://dl.dropbox.com/u/17111233/Roll_up.rar

Colour safe chart:

jamespetts

The problem with weight classes is that one creates an entirely new layer of abstraction to deal with weight when the actual figure for weight is already used. If weight classes apply to vehicles, then there does not seem to be any advantage over the present system  (as it still will not have the desired effect of differentiating between the weight necessary for ways and the weight necessary for bridges). If they apply to convoys as a whole, then they have the same difficulties as using actual total convoy weight would apply.
Download Simutrans-Extended.

Want to help with development? See here for things to do for coding, and here for information on how to make graphics/objects.

Follow Simutrans-Extended on Facebook.

merry

Now this is a subject that has - in a related way - recently come to the fore in my gradual play through Pak Britain Experimental timeline.
I have found that by the 1960s I cannot realistically construct relatively slow speed fregiht branches, because the weight and speed increases are too closely related. Actually, I could, but the slow 150t track is more costly to maintain than (concrete sleeper) fast 150t track. So there is little point.

Considering the cost of track - there are many contributing factors. A route may be built for heavier or lighter traffic (which is absolutely axle-load related). It may also be built for fast, or slow traffic - the hammer action of fast traffic makes maintenance and initial cost much higher. And there is a big difference between track for freight-only or passenger & freight use, becuase the safety requirement for freight is very low compared to passengers: pointwork requires facing locks for passengers (ie signalling is costly), the quality of track line and level must be higher, etc. Passenger trains can pass over freight track but only at low speed and not in normal circumstances.

The axle load issue is very relevant for convoys...many of the UK diesel locos used in the last 40 years are found across the widest reaches of the uk network, but in Simutrans they are almost all over 105t weight, so can only go on the heaviest (and fast) track. In reality the class 37 could go most places, indeed has the same route availability as a class 20, 31, or 55 whilst being much heavier & more powerful. But in simutrans...well, try using anything but a class 20 on 88t track (the only useable low-cost branch track in 1965!)
I did a quick check:
class 20 is 73t on 4 axles: 18.3t/axle
class 37 is 107t on 6 axles: 17.8t/axle (some variants 102t ie 17t/axle)
class 55 is 101t on 6 axles: 16.8t/axle
class 40 is 135t on 8 axles: 16.8t/axle (it's a 1Co-Co1 wheelset)
In ST, the class 20 can use lightest  track and the class 40 only the heaviest, while IRL it's the other way round. Go figure?!?
This is why BR (and in one way or another their predecessors) used route availability to avoid the complicated confusion that otherwise results!

Now, I agree that bridges are total weight dependent, but we can't use the whole convoy as long convoys on short bridges would distort the behaviour. Long bridges have a different cost to short ones per tile, so some of this is included in the price. However, the 'point loading' of one vehicle is also a key factor in bridge dynamics, and so I propose for simplicity the weight limit is that of the heaviest vehicle in the convoy. It's close enough to relality to be good.
Another point: some time ago, it was mentioned that if a weight limit were exceeded and no viable alternate route existed, then (up to some limit) it would be OK to traverse track substandard for a convoy (say a low-weight-limit bridge) at low speed. This is highly realistic, but doesn't seem to occur in ST at the moment.
And an example from reality: A King locomotive can traverse Victoria bridge (on SVR near Bewdley) but only at walking pace, and double-headed trains must uncouple the leading loco which crosses separately with the train, and the convoy recombines after. Very slow!

So I would suggest a few things:
* there needs to be another factor in track selection, that of freight/passenger use
* bridges should remain dependent on greatest vehicle weight per convoy
* track should be axle-load dependent
* if a convoy needs to traverse lower-weight track, or passenger trains over freight track, because it's the only route, then it can do so at much reduced speed, and a warning would be generated. There might even be a cost penalty for the operational complexity (now that'd be realistic).

Ooh, a lot here. I hope it helps the discussion. I can see there are some big-ish changes.
If only one thing were taken, I think it's far more important the track weight limit being axle load not vehicle weight.
TTFN
merry

jamespetts

Hmm, this is very interesting. Thank you for the contribution! I know that the Standard developers were planning to add axle loading limits at one point, and made some preparations in the code for it, but no work seems to have been done on this recently. I had been waiting to see what happens with this before considering this issue in detail again.

There are two separate issues here: (1) the setup of the track (which tracks have what maintenance costs, speeds and maximum weights, and also what sorts of tracks that there are); and (2) the suggestion of additional features in the code.

The current short-term planning means that item (1) is currently under review, but item (2) will take some time to be ready to be done, if it is considered advisable. The current development plan is to iron out the existing bugs in the -devel branch and finish testing the new physics features before releasing a new version. Whilst work is done by Bernd on fixing the physics bugs, I have been concentrating on pakset work, including:
(1) integrating the new bridges into the Experimental pakset;
(2) re-organising the different sorts of railway track a little, and making sure that there are distinct graphics for each sort of track, as well as making the selection more realistic; and
(3) adding some new vehicles, including a better selection of narrow gauge vehicles.

Once the next version is released, the programming priority will be to progress with some of the economic features discussed here (especially those marked "balance critical"). The first of those items that I plan to undertake will be that set of features relating to vehicles (usage based maintenance costs, overhauls, secondhand purchases and easier replacing). Once that is done, I  plan to focus on way based economics, including use based way maintenance costs, mothballing, way upgrading costs and way renewal.

That phase of focussing on the economics of ways will be the time when the question of how to deal with weight limits will need to be finalised. Discussion as to how best to deal with it is useful in the meantime. However, of more use in the short term would be any thoughts on how to get a more realistic selection of track types throughout the eras (and once the types are defined, new properties such as axle loadings and usage based costs can later be added as, when and if those features are implemented).

One suggestion that does interest me is using vehicle weight for bridges, which might be a workable solution, if we can think of a sufficiently clear means to represent that in the UI. I should be interested in particular in people's views on:
(1) whether using vehicle weight for bridges is a good workaround;
(2) if it is, how best to explain to players using the UI that this is what is happening; and
(3) the question of axle weights (and how best to represent these in the UI).
Download Simutrans-Extended.

Want to help with development? See here for things to do for coding, and here for information on how to make graphics/objects.

Follow Simutrans-Extended on Facebook.

kierongreen

Class 20, 37 and 55 have RA 5, Class 40 have RA 6...

sdog

To answer your second question
Quote(2) if it is, how best to explain to players using the UI that this is what is happening;

this quote from merry helps:

QuoteAnother point: some time ago, it was mentioned that if a weight limit were exceeded and no viable alternate route existed, then (up to some limit) it would be OK to traverse track substandard for a convoy (say a low-weight-limit bridge) at low speed. This is highly realistic, but doesn't seem to occur in ST at the moment.
And an example from reality: A King locomotive can traverse Victoria bridge (on SVR near Bewdley) but only at walking pace, and double-headed trains must uncouple the leading loco which crosses separately with the train, and the convoy recombines after. Very slow!

The current way of convoys not finding a route is very intransparent and confusing. Slow speed at insufficiently capable route however would quickly allow players to learn how the weight limits work, by trying. Having messages at two points, will be necessary: A warning in the notification window, and more importantly a longer explanation at the convoy dialogue. (the first thing i'd do when i see an extremly slow train is clicking on it.*


*an explanation for the 4 km/h for convoys with insufficient power should be helpful too.

ӔO

In NA, since there are only a few types of locomotives that all fit within the defined axle load, they use a system of track classes and plates for maximum loading gauge.

http://en.wikipedia.org/wiki/Track_class_%28United_States%29#Track_classes
http://en.wikipedia.org/wiki/Loading_gauge#North_America

In a game, to the player, it would be easier if all of this was just represented as one value.
My Sketchup open project sources
various projects rolled up: http://dl.dropbox.com/u/17111233/Roll_up.rar

Colour safe chart:

jamespetts

What do you mean, exactly - all of what represented as one value, and what would that value be? Remember, this system has to work for roads as well as railways...
Download Simutrans-Extended.

Want to help with development? See here for things to do for coding, and here for information on how to make graphics/objects.

Follow Simutrans-Extended on Facebook.

sdog

Quote from: jamespetts on May 31, 2012, 04:33:44 PM
Remember, this system has to work for roads as well as railways...

Why? The total weight of a convoy is enough for roads. Different waytypes could very well have different systems.

jamespetts

Different waytypes having different systems would introduce enormous complexity both for the player and in the code, and really should be avoided if at all possible. Also - why should roads be based on the convoy weight rather than the vehicle weight?
Download Simutrans-Extended.

Want to help with development? See here for things to do for coding, and here for information on how to make graphics/objects.

Follow Simutrans-Extended on Facebook.

ӔO

Quote from: jamespetts on May 31, 2012, 04:33:44 PM
What do you mean, exactly - all of what represented as one value, and what would that value be? Remember, this system has to work for roads as well as railways...

I mean that axle load + convoy mass would be unnecessarily complex.



Things are a lot easier for road vehicles, actually. Nearly all road vehicles are built within the specifications defined by the government. They all must comply with GVWR and GCWR laws. As long as they are within the GVWR and GCWR limits, they will never exceed maximum allowed axle loads for roads or bridges. Excepting older bridges, which were never designed with the newer road laws in mind.
My Sketchup open project sources
various projects rolled up: http://dl.dropbox.com/u/17111233/Roll_up.rar

Colour safe chart:

jamespetts

Quote from: ӔO on May 31, 2012, 09:41:32 PM
Things are a lot easier for road vehicles, actually. Nearly all road vehicles are built within the specifications defined by the government. They all must comply with GVWR and GCWR laws. As long as they are within the GVWR and GCWR limits, they will never exceed maximum allowed axle loads for roads or bridges. Excepting older bridges, which were never designed with the newer road laws in mind.

Except that there are rather a lot of older bridges for roads with all sorts of weight limits all over the place...
Download Simutrans-Extended.

Want to help with development? See here for things to do for coding, and here for information on how to make graphics/objects.

Follow Simutrans-Extended on Facebook.

wlindley

Isn't it true that, from the driver's perspective, he sees only a sign "Load Limit X" and he knows whether "X" is greater or less than his vehicle's weight limit.

So long as that one number X is displayed in the Depot purchasing dialogue and when the user clicks on a convoy, surely it does not matter what data or calculations are in the dat, pak, or program? 

I note that as you add vehicles to a convoy, depending on the internal algorithms, the displayed number X may both increase and decrease.

sdog

Axle load doesn't matter for roads. Total weight is the limiting factor. Axle load is just required not to exceed acceptable values, by having an appropriate number of axles. Limits on ways are total weight, determined mostly by bridges. The speed limits already take care of weight sufficiently. You can use the maccadam road at the low speed limit with a 40t articulated lorry as well as with a car.

For rail axle load of the engines is a relevant factor though, they are considerably denser than road vehicles.

It doesn't apply at all to other waytypes. For airports the weight is relevant with the runway length mostly. (at experimental scale this does matter, unlike standard with 1k tile length, we need at least 3 km of runway!)

For ships and canal boats it does not matter at all. (displacement of water)

For maglev it likely is not required either, it's PAX only anyway and the distributed traction means there is no part really heavy, (as with modern highspeed trains).




you (and players) would have just two systems to distinguish:


for rail the "route availability" (include bridges! assume if it has the same route availability class, it will also take the total weight. anything else would be silly engineering, we need not simulate)


for road it is just total weight of the convoy [a bit boring as it ends at 40t (it's the same in britain, isn't it)] with a little application in old bridges.

ӔO

#18
Axle load and wheel load is a factor for roads, but the end user will never have to worry about it. This is because the end users only have to worry about maximum payload and placement of mass. The maximum allowed payload is usually printed somewhere on the vehicle because the manufacturer has already designed their vehicles to fit within the limits with the maximum limitations in mind. If the vehicle was only designed to have a GVWR of 20t, the number of wheels and axles used will match accordingly. For 40t, the axle and wheels will be increased accordingly.

Here is an example: http://translate.google.com/translate?sl=ja&tl=en&js=n&prev=_t&hl=en&ie=UTF-8&layout=2&eotf=1&u=http%3A%2F%2Fblog.dosekan.com%2F%3Feid%3D515604&act=url


What makes rail difficult, is that they were once private infrastructures, unlike roads, which are public. Private infrastructure is, normally, only built with the company's fleet in mind, so it may not be compatible with another company's fleet. Also, the concentration of mass can fluctuate a lot, because the maximum axle load is very high, compared to road.

As an example, you can have a 100t for Bo-Bo, or 150t for Co-Co, if the maximum axle load is 25t. For trucks, with a 10t maxmium, at most, you will have a 20t to 30t jump, by going from 2 axles to 3. They are both a 50% jump in mass, but the scale of the jump is not the same.
My Sketchup open project sources
various projects rolled up: http://dl.dropbox.com/u/17111233/Roll_up.rar

Colour safe chart: