News:

Want to praise Simutrans?
Your feedback is important for us ;D.

Bridges

Started by The Hood, December 10, 2011, 10:31:24 AM

Previous topic - Next topic

0 Members and 3 Guests are viewing this topic.

The Hood

There is an absolute limit imposed on all bridges of +7 height. This can be reduced for individual types.

ӔO

From my testing, the maximum height is 8 and the shortest span possible with that height difference is 9 tiles.
which makes me wonder if a dat specified max height is necessary at all.
My Sketchup open project sources
various projects rolled up: http://dl.dropbox.com/u/17111233/Roll_up.rar

Colour safe chart:

jamespetts

That is interesting. Would it be possible to give a chart of minimum length per tiles of height to enable the issue to be considered more fully?
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

with elevated ways being buildable on water now, is it even necessary to keep the viaducts as bridges?


modification: to answer my own question: elevated ways can be only one level high. another level would require the player to stack them, quite awkward.


This is also a real construction option, with a bridge between two sections of elevated way, and artificial slopes, one can easily build a one square wide three level deep drop to bridge.

prissi

You need a bridge to connect to an elevated way ... but you can keep it of course to single tile length.

sdog

Ah, yes. I just realized that what i had in mind is not possible at all at this time.

A bridge can be connected to an elevated way only at one end. It can not be used between two pieces of elevated way.
If one needs a bridge

example (E=elevated way, B=bridge, _ = any ground)

___EEE
this works
__BEEE

EEE__EEE
here it doesn't
EEEBBEEE

btw, it is possible to attach elevated ways directly to artificial slopes.  That's also whats used to build diagonal "bridges" (workaround).

ӔO

Quote from: jamespetts on January 10, 2012, 11:35:49 PM
That is interesting. Would it be possible to give a chart of minimum length per tiles of height to enable the issue to be considered more fully?
if each step is allowed to have a 2 tile height difference, then these would be the numbers.
H= height
L= length end to end.

H:8, L:9
H:6, L:7
H:4, L:5
H:2, L:3
My Sketchup open project sources
various projects rolled up: http://dl.dropbox.com/u/17111233/Roll_up.rar

Colour safe chart:

wlindley

Loving the new bridges, but could we go back to the issue of building cost and operating cost?  All the costs seem much too low.

For example, Phoenix Arizona in 2008 opened its new tram system; the cost for the entire 20-mile line, 50 trams, and a full modern maintenance facility was $1,200,000,000. The cost of one mile of line laid in the street was about $20,000,000. The single main-line 500 metre bridge over the Salt River in Tempe, however, cost $300,000,000. 

In general, bridges have cost about 10 times at-grade rail.  A good range would be from 5 times to 20 times depending on bridge type.

As far as game play -- bridges are some of the most spectacular parts of our railroads, and deserve the best planning.  A higher initial cost would reflect this.

jamespetts

WLindley,

I am very interested in real world pricing information for balancing purposes; is your "in general" figure based just on the two examples that you gave, or is it based on anything else? And has that ratio remained largely constant throughout history, or was it different, say, 100 or 150 years ago?
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

I know I've read a report before on how bridge prices always end up being about 30% higher compared to initial estimates 60% of the time.
it might have been on a news piece about the costs of the new oakland bay bridge in san fransico.
My Sketchup open project sources
various projects rolled up: http://dl.dropbox.com/u/17111233/Roll_up.rar

Colour safe chart:

wlindley

I found a reference of page 53 of the Google scan of the 1832 American Rail-Road Journal for construction costs on the Fifth Division Baltimore & Ohio Railway.

This division had 11 miles of grade, and "three bridges of one arch each, and of the following chords, to wit, 30, 20, and 10 feet respectively; and one viaduct of the Rail-road for the Georgetown and Frederick turnpike road, of stone abutments and superstructure of wood of 24 foot span. This viaduct is elevated [16 feet 1.2 inches] above the gradated surface of the Rail-road."

The 11 miles of gradation cost $66,614 (about $6,000 per mile) while the masonry for division altogether cost $12,068 (for 84 feet (!) in four bridges plus an unspecified number of smaller culverts). 

Altogether my feeling is that this example suggests a Simutrans tile of masonry bridge should cost about 20 times a standard tile. 

This document for Washington State's Sound Transit lists a cost of $10,140 per foot for a long post-and-beam bridge, while this document from the State of Michigan lists an installed cost for railroad sidings (spurs) to be about $170 per foot (for 115-pound rail with 9-foot concrete roadbed)... the bridge costing sixty times a simple siding. However a siding costs less than a main-line track... again, 20 times seems about the right answer.

p.s., That 1832 article has a variety of figures quoted for European railways as well, unfortunately the scan process has obscured too many of them.

jamespetts

Very interesting, thank you!
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.

AP

If the cost of bridges is being seriously reviewed, it might be worthwhile to revisit the discussion about the cost of altering terrain (link), to ensure the desired incentives and balances are maintained.

The Hood

So where does this leave us in terms of costs for bridges and altering terrain? Looking at rail only for the time being:

Track Costs (maintenance in brackets):
wrought iron: 30.00 (3.00) for 60km/h
improved wrought iron: 150.00 (16.00) for 110km/h
WSSR: 250.00 (26.00) for 145km/h
WSSRI: 400.00 (36.00) for 175km/h
CSSR: 500.00 (44.00) for 200km/h
CSSRI: 600.00 (48.00) for 225km/h
High Speed: 1,000.00 (72.00) for 320km/h

1) are these costs acceptable (a) relative to one another and (b) as an absolute reference point for the game?

Land costs:
alter land: 1250.00 (0.00)
artificial slopes: 1500.00 (0.00)

2) Already discussed in linked thread, but (a) what balance should these have relative to one another and (b) what should they cost relative to tracks?

Bridge costs:
Wooden trestle: 250.00 (7.00) for 30km/h
Brick Viaduct: 1,900.00 (72.00) for 160km/h

Elevated ways:
Brick Arch elevated way: 250.00 (26.00) for 145km/h

3) what costs should bridges have relative to altering terrain and altering tracks (NB these are not independent: fixed by whatever gets decided about terraforming and track costs)?

Other points to note:
- Partially the problems are one of scale. Taking the "official" scale of 1 tile = 1kmx1km (which I've never liked even remotely because of the graphical absurdity of it all) then steep embankments/cuttings are the way to go - the tapering is irrelevent on that scale. The graphical scale however is that 1 tile is approx 30-100m (depending on context), which probably makes embankments/cuttings with a slope slightly more realistic, but wouldn't entirely rule out the sheer drop type.
- If "sheer drop" is considered as in the suggestion in the other thread, i.e. make it very expensive as it's unrealistic, then what about elevated ways? These can only be built one high, but are effectively a "sheer drop" plus track and plus maintenance. Why ever bother with these except where requiring to cross another way? Is that the right way to go? Should they be cheaper or more expensive than sheer drop terraforming?



AP

#119
My position is that the pricing structure and balancing should encourage a creative mix of all the options to be used by the player, to maximise price advantage, and thereby appear realistic.

As such, building on that earlier work, I think the balance needs to be:

( 2x cost_alter_land )  < all bridge prices (possibly except slow timber bridge) < cost_set_slope < elevated way


       
  • 2x cost_alter_land is the cost of raising one tile (assuming infinitely long embankment)
  • I assume bridges would be priced in a spectrum, like track, to represent their speed restrictions etc?

       
  • Elevated ways have two quite distinct advantages over embankments, as I see it, hence are very useful:

       
  • Flexibility - unlike bridges they can be built diagonally, and can have intermediate junctions.
  • Impact - unlike sheer embankments, they preserve buildings and other urban infrastructure underneath,  (and preserves passenger-generating buildings). This results in a very considerable saving on urban demolition (which can easily exceed $10000/tile) compared to a sheer embankment. As such I think the per-tile pricing for elevated ways should be greater than the price for cost_set_slope. It could take into account the cost of a typical urban building (obviously not the very expensive ones which are to tall to bridge over). Possibly maintenance of elevated ways should still match the brick viaduct though.
Quotealtering terrain and altering tracks (NB these are not independent: fixed by whatever gets decided about terraforming and track costs)?
Can you clarify this? - which variable is controlling the cost for track demolition/changes then? Because track changes at present are far cheaper than any terrain alterations...

Edit - add the bullet for 2x cost_alter_land

The Hood

Quote from: AP on January 21, 2012, 11:00:31 AM
Can you clarify this? - which variable is controlling the cost for track demolition/changes then? Because track changes at present are far cheaper than any terrain alterations...

I meant that if you decide what the ratio of costs track:terraforming is and track:bridges you automatically define bridges:terraforming.

AP

Ah, I see, quite true! I'm not sure track:terraforming is as important as the other two in terms of influencing player choice of route/costruction, i'd let that be the one to be defined 'automatically'.

The Hood

More thoughts:
- I don't want to set the maintenance of bridges punishingly high, especially if they are to cost e.g 20x normal track. How about double the cost of track of a similar speed?
- elevated ways shouldn't cost more than bridges per tile - they are effectively bridges that can go round corners
- why should bridges be cheaper than retaining walls and earth that we are envisaging for artificial slopes? After all they have disadvantages that AP has suggested e.g. demolishing buildings in their way, and don't include track either...

Anyway, keeping the cost of rails fixed for now, and merging in the ideas so far, I suggest the following for comment:

Brick Viaduct/Elevated way: 6,000.00 (72.00) for 160km/h, unlimited length and height
Wooden trestle: kept as is at 250.00 (7.00) for 30km/h given the very low speed limit. Unlimited length but height limited to +3 or 4?
Proposals for new bridges:
Brick Viaduct 115: 5,000.00 (50.00) for 115km/h, unlimited length and height limited to +5
Masonry Viaduct (not drawn yet): 3,500.00 (30.00) for 70-80km/h. Unlimited length but height limited to +4 (mid weight limit)
Wooden trestle stone viaduct: 3,000.00 (40.00) for 70km/h Unlimited length and height (low weight limit)
Wrought Iron lattice viaduct: 3,500.00 (80.00) for 130km/h Unlimited length and height
Wrought Iron lattice girder: 5,000 (60.00) for 145km/h max height +2
Tubular box girder: 4,000 (50.00) for 120km/h max height +5
Steel beam: 3,500.00 (50.00) for 120km/h max height +3 moderate weight limit max length 10
Steel Truss: 6,500.00 (90.00) for 175 km/h - max height +2 (mod-high weight limit) max length 10
Wrought Iron Arch: 5,500.00 (60.00) for 145km/h max height +3
Wrought Iron Bowstring girder: 5,000.00 (55.00) for 145km/h max height +2
Steel/Concrete Bowstring girder (not drawn yet): 6,000.00 (64.00) for 160km/h max height +3
Plate Girder Brick: 5,000.00 (50.00) for 145 km/h - max height +1 max length 4
Plate Girder Concrete: 8,000 (72.00) for 175km/h - max height +1 max length 4
Concrete Viaduct: 10,000.00 (80.00) for 175 km/h no limits
Concrete Viaduct: 12,000.00 (100.00) for 225 km/h no limits
Concrete Viaduct: 20,000.00 (150.00) for 320km/h no limits

Alter slope tool: 1,000.00 (0.00) (this will rack up though with extra height...)
Artificial slope tool: 2,000.00 (0.00)

Thoughts?


AP

QuoteI don't want to set the maintenance of bridges punishingly high, especially if they are to cost e.g 20x normal track. How about double the cost of track of a similar speed?
I agree. Maybe a little low, but perhaps best reviewed in play-testing.

Quote from: The Hood on January 21, 2012, 05:00:02 PM
- why should bridges be cheaper than retaining walls and earth that we are envisaging for artificial slopes? After all they have disadvantages that AP has suggested e.g. demolishing buildings in their way, and don't include track either...
Hmm.  They shouldn't be cheaper than 'simple' earthworks, but should be preferred over 'complex' ones. The trouble is the limited number of variables. At what point bridges are preferred over earthworks is, I think, the more complex question.

QuoteAlter slope tool [cost_alter_land]: 1,000.00 (0.00) (this will rack up though with extra height...)
Artificial slope tool [cost_set_slope]: 2,000.00 (0.00

In that scenario, across open fields, a +1 sheer embankment costs the same as a sloped embankment. And also it's almost always cheaper to build a +2 sheer embankment (2000/tile) than a bridge, though the former look really crude in game (not an insignificant consideration) and have 0 maintenance excl track, whereas a +2 sloped embankment (more attractive) costs 6000/tile when it ought to be cheaper.

If a sheer embankment is equally cheap as a sloped embankment, nobody will build the sloped embankments even when they have the space, so the sheer one has to be at-least-slightly pricier.
To rectify that for +1 needs cost_set_slope > 2x cost_alter_land
To rectify that for +1 and +2 needs cost_set_slope > 6x cost_alter_land (if I got my maths right...)






jamespetts

Do we not need to look into the historical costs of bridge maintenance?
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.

The Hood

So if I understand you correctly, we could set artificial slopes at 6,000. This would make them as expensive as bridges for forming embankments, but would have an application in situations where you needed to create a ledge to fit the track along without demolishing an entire mountain. Probably about right.

Regarding historical costs of bridge maintenance, be my guest in experimental, but as a starting point I would just get something that allows each bridge type to have a niche in game, especially in standard where the number of parameters to play with are reduced. Don't forget there's no inflation, and no increase in maintenance with age. Keep it simple...

AP

QuoteWe could set artificial slopes at 6,000
Yes, that would seem a good value to use initially. Higher would have advantages as stated previously, but possibly also disadvantages.

QuoteRegarding historical costs of bridge maintenance, be my guest in experimental, but as a starting point I would just get something that allows each bridge type to have a niche in game, especially in standard where the number of parameters to play with are reduced. Don't forget there's no inflation, and no increase in maintenance with age. Keep it simple...

I agree I think - since the graphics now exist and are going into the game, best to get the values/parameters roughly balanced, then maybe review later if better information becomes available.

wlindley

To keep things simple:

       
  • Purchase cost should increase (slightly exponentially) less than the square of maximum weight load (which is a game factor in Experimental) and maximum length (for bridges where length is a factor).
  • Maintenance cost should increase (slightly exponentially) less than the square of maximum speed
In each era a small variety of bridges should be available:

       
  • slow, light: cheap to build and maintain.
  • slow, heavy capacity: moderate to build, cheap to maintain
  • fast, medium capacity: fairly cheap to build, fairly expensive to maintain
  • fast, heavy capacity: fairly expensive to build, fairly expensive to maintain
The goal here is to force the player to make decisions, highlighting each player's and each terrain's unique economics and demand in engineering techniques.

AP

Sounds good.

QuoteThe goal here is to force the player to make decisions, highlighting each player's and each terrain's unique economics and demand in engineering techniques.
I agree entirely.

The Hood

Would either of you like to make any suggestions as to which bridges fit which and put some numbers to them? That would help me massively...

Spike

Quote from: wlindley on January 12, 2012, 12:36:19 AM
Altogether my feeling is that this example suggests a Simutrans tile of masonry bridge should cost about 20 times a standard tile. 

Thank you, this is a very useful rule of thumb!

wlindley

I have assembled a spreadsheet here where the second page is based on data from the Experimental .dat files, and a set of linear and logarithmic factors.  The model is calibrated so the early wooden and brick bridges approximate their current costs.  The later faster and higher-capacity bridges are generally more expensive than currently, although note: The Concrete Spanning Bridge (Light) would now cost half as much to build while maintenance costs would double; meanwhile the Heavy version would cost more than twice as much to build, and maintenance costs would also more than double.  Available during the same era, however, are several bridges like the Steel Box Girder which are a significant savings in both building cost and maintenance, if slightly lower capacities and speeds are not an issue.

Feel free to adjust the factors at the top of the page.

jamespetts

Thank you - that looks very useful! I shall have to look into that in more detail when I get the time.
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.

The Hood

Wlindley

A few qustions about this:
1) does this incorporate the 20x standard tile rule of thumb? Prices generally seem quite low compared to what I worked out above, and seem a bit low compared to the raising land option discussed above.
2) This is for experimental, which for historical reasons has separate bridge types (and I don't think jamespetts is planning on changing them to the types I have drawn which will be the ones in standard). How hard would it be for you to do a similar job for standard?

jamespetts

I have yet to decide exactly how to integrate the latest Standard bridges into Experimental, but, whatever I do, I shall certainly be making use one way or another of all of The Hood's excellent bridge graphics.
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

#135
Ah, I used Experimental's data files because they had the weight limits.  Here is a revised spreadsheet, with eight pages (four for Standard, four for Experimental), (track, bridge, construction, and maintenance costs for each).  The Standard base sheets do some weight lookups into the data from Experimental. 

I hope you find the construction and maintenance costs here to have sensible ratios to the costs for similar at-grade track.  If not, adjust the parameters in the yellow shaded boxes.  Each Cost and Maintenance line has a Subjective entry field as well, in which I have placed some "fudge values" to make the very early and very late bridges seem right.

For Standard, a difficulty arises in that the two Brick viaducts, which are available quite early, have much higher speed limits than will be possible for years after their introduction, which skews the costs.  Perhaps a slightly wider variety across the years would be appropriate.

The Hood

OK,  I've now added the bridges (along with a couple of extra ones) and rebalanced the costs. Do let me know how these costs work for you.

wlindley

Very nice!  And the additions are simply grand. This is a nice variety, with a little emphasis on economic as well as aesthetic influence on the game. I quite like the wooden trestles with stone pillars, and the iron lattice.

Should I do the same spreadsheets for roads now? =grin=

The Hood


wlindley

Standard and Experimental road bridges, with possible construction and maintenance costs, here.  As before, the yellow squares are for input coefficients. 

For Standard, the early bridges are just about right although MasonryViaductCobbles probably should have its construction cost increased.  The concrete bridges all should have their cost increased by five or six times... other than that, things are fairly sensible as-is.

Experimental's curves are much sharper, and it seems the bridges need much more adjusting.