News:

Simutrans Tools
Know our tools that can help you to create add-ons, install and customize Simutrans.

Signal Boxes/Signalling Centres should have unlimited signals

Started by fam621, November 12, 2017, 12:20:11 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

fam621

I have just finished playing Simutrans for now and I remade a station and remodelled a complex junction and what I want to know is that why does the large ERTMS Signalling Centre have only a max capacity of 175 signals. Like modern day signalling centres irl don't have max capacity. Only a question but can the ERTMS Signalling Centre have unlimited signals please?

Edit: Change of topic name

jamespetts

The idea of a maximum number of signals is that a signalbox has a finite capacity to control things. There is no easy way to code a finite capacity for controlling points, so the only variables that we can usefully use for a finite capacity are area or the number of signals.

Do you have data to suggest that the capacity of the ERTMS Centre as currently set is incorrect?
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.

fam621

#2
There is no date available on how much signals can be controlled but how can you increase signalling control?

jamespetts

I do not understand that at all, I am afraid. Can you explain more clearly?
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.

fam621

What I mean is that when I looked up on Google 'how much signals can a signalling center control' there is nothing which says how much a signalling center can control.

Rollmaterial

In Germany, operations control is being gradually concentrated in 7 operations centres (Betriebszentralen). These arguably control several thousands of signals each.

jamespetts

So there are no data to suggest that there is anything wrong with the current figures?

Edit: Rollmaterial - the limit is needed to simulate the reason that there are 7, rather than just one in examples such as that.
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.

fam621

I'm afraid so James but how can you edit it so that you can be allowed to control more signals?

jamespetts

I am more interested in why one would do so and to what number that one would edit it.
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.

fam621

Because, on my fictional WCML and (6 or more signals) on the ECML, there are not enough signals to cover both routes (south of the border) and I think that me asking how to edit how much signals a control center could handle would be beneficial for me so I can add those missing signals. I would edit it to 200 signals.

Rollmaterial

By editing the relevant .dat file and compiling it, quite simply.

fam621

Ok. How do you comply after you edit the .dat file.

jamespetts

Oh, you want to edit it yourself? I cannot remember the name of the parameter in the .dat files, but there is a .dat file reference on the forum somewhere that should tell you all about the various parameters for signalboxes. How to compile the pakset once you have made your changes is already explained in the tutorial on how to make additional liveries: the principle is the same.

In terms of the broader question of how the signal limits ought to be calibrated, it is not much use thinking about how many signals that an actual signalbox of the relevant type would control, as a real signalbox would control level crossing signals, shunting signals, repeaters, yard signals and many other things that are not directly simulated in Simutrans. The signal limit is instead intended to be proxy for an overall limit on scope of network that a signalbox can handle. This is much easier to calibrate for smaller signalboxes than larger signalboxes.

Thus, whilst it does not make sense to abolish the limit entirely as it has a specific purpose, if there be some reason to believe that the limit is too low and that some other specific higher number would better reflect how these sorts of signalling centres are used in reality, then I am certainly amenable to amending the limit - but cannot sensibly consider that without more detailed information as to the reason that the current limit of 175 signals (which is a lot for Simutrans) is too low. For example, if a real ERTMS signalling centre covers network of scope and complexity that is considerably greater than that which can be covered by one in game, then it would be reasonable in those circumstances to amend the signal limit to allow an ERTMS centre to control the same scope of network as it would be able to control in reality.

If anyone has any clear data on this, that would be most helpful. Otherwise, you might want to consider either using more ERTMS centres or an ERTMS centre combined with some smaller signalling facilities.
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.

DrSuperGood

Modern control centres theoretically have no limits to the number of signals they can control thanks to computer technology. One could likely engineer one capable of managing hundreds of millions of signals if the need arises. What is more a limiting factor is what is it practical for a single centre to control.

The reason multiple centres are used in real life is for redundancy and failure limitation. Having a single control centre means a single point of failure for the entire network, similar to the star network pattern. By breaking control down into separate areas, possibly even with added redundancy, the maximum effect of a single point of failure would be the failure of only 1 region and not the entire network. Seeing how signal failure not only causes huge costly delays to train operators but also puts lives at risk it is well worth the added overhead of multiple centres to manage the network, even if the technology exists to control all signals in the entire network from a single centre.

Instead of defining some arbitrary limit, it would be better to define other limiting factors that force the construction and maintenance of multiple control centres. For example greater than 90% of signals controlled by control centres must be within 400 (or some number) tiles of a control centre, and that there must be at least 2 (or some number) control centres before any will work.

Another approach would be to have no limits but rather add a per signal cost to the centre, like there is in real life. The more a single centre controls, the higher its maintenance is. This could be applied per signal, or to the centre itself.

fam621

You know what, I'm going to put control centres in different places line irl now. :)

jamespetts

Adding different kinds of limits will take coding resources that are needed elsewhere, and for only very minor benefit. I should note that there is already a per signal maintenance cost.
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.

Vladki

Just for comparison. There are only 2 control centers for the whole Czech Republic. I do not think they can serve as backup for each other. However they do not control the whole network. There are still man lines that are controlled locally. But the plan is to do it.

But I have no idea how many signals are on the czech railway network.

Sent from my ONEPLUS A3003 using Tapatalk


jamespetts

It is really a matter of finding a sensible signal limit that reflects how large a part of a network that it is sensible for one centre to control.
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.

Rollmaterial

I believe one signal per kilometre of track would do for a sensible approximation, as it is the order of magnitude of the distance between a stop signal and its distant signal.

jamespetts

How could that be used to set a limit on the total number of signals?
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.

DrSuperGood

Maybe it is a usability problem? Perhaps the limit should be made globally where the maximum is defined by the number of centres and the current is total number of signals. Either the original topic creator was really fussy about maintenance costs, or they were complaining that simply building another centre was not a simple solution in this case.

fam621

#21
Quote from: DrSuperGood on November 12, 2017, 07:33:42 PM
Modern control centres theoretically have no limits to the number of signals they can control thanks to computer technology. One could likely engineer one capable of managing hundreds of millions of signals if the need arises. What is more a limiting factor is what is it practical for a single centre to control.

The reason multiple centres are used in real life is for redundancy and failure limitation. Having a single control centre means a single point of failure for the entire network, similar to the star network pattern. By breaking control down into separate areas, possibly even with added redundancy, the maximum effect of a single point of failure would be the failure of only 1 region and not the entire network. Seeing how signal failure not only causes huge costly delays to train operators but also puts lives at risk it is well worth the added overhead of multiple centres to manage the network, even if the technology exists to control all signals in the entire network from a single centre.

Instead of defining some arbitrary limit, it would be better to define other limiting factors that force the construction and maintenance of multiple control centres. For example greater than 90% of signals controlled by control centres must be within 400 (or some number) tiles of a control centre, and that there must be at least 2 (or some number) control centres before any will work.

Another approach would be to have no limits but rather add a per signal cost to the centre, like there is in real life. The more a single centre controls, the higher its maintenance is. This could be applied per signal, or to the centre itself.

Very true and I would like for it to be introduced to SImutrans Extended but of course priorities count! But I am not rushing for it to be added but still

fam621

I think that maybe for Signal Boxes/Signalling Centres having unlimited signals, add a per signal cost to the centre, like there is in real life.

jamespetts

Quote from: mammothim2906 on November 11, 2018, 10:38:01 AM
I think that maybe for Signal Boxes/Signalling Centres having unlimited signals, add a per signal cost to the centre, like there is in real life.

Each signal has its own maintenance cost, so adding a maintenance cost to the signalling centre would be unnecessary and confusing.

As to limits, there needs to be a realistic incentive to build a realistic number of signalling centres, or else no player will ever build more than one for the whole world. That will be a balance problem, as the signalling cost will then be unrealistically low.
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.

DrSuperGood

In real life the number of centres is determined by the need for redundancy as well as failure mitigation. In Simutrans your signalling centre will not burn down randomly or be destroyed by an Earthquake like they can be in real life. Hence there is no concern for redundancy or failure mitigation in Simutrans.

One could simulate this by forced mechanics. For example one needs at least 2 signalling centres for any signal to start working, and then every 1,000 odd signals an extra signalling centre. One can mention in documentation or in help that this limit is to reflect the need for redundancy.

jamespetts

These extra mechanics would require substantial extra coding and are a very low priority compared to several years' worth of additional features and bug fixes currently queued.

Also, having a separate redundancy mechanic for these signalling centres would make it much more difficult for players to understand the parameters and limitations than if the signalling centres had the same signalbox limitation mechanics as all other siganalboxes.
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.

fam621

Ah, I can fully now see your point now James (and those other people who have commented on this topic).