Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Jeoshua]

Well to me, that's one of those "probably should be a concern in the first place" things.

In most cases, building supports or leaving a few load-bearing walls would be good enough to keep a fortress from collapsing.  In general it would be required of the player to be a bit smarter about load bearing structures, but if support can extend vertically and horizontally then it should be fairly logical.

I can imagine a fortress design that looks roughly like this:

Code: [Select]

xxxxxxxxxxxxx
x...x...x...x
x.....o.....x
x...x...x...x
xx.xx...xx.xx
x...........x
x.o..<o>..o.x
x...........x
xx.xx...xx.xx
x...x...x...x
x.....o.....x
x...x...x...x
xxxxxxxxxxxxx

The 'o's would be supports.  They would bear a lot of the load for the interior of the large central room.  In addition, if this design were stacked vertically it would have interior load bearing walls all over the place.  So really it doesn't require a whole lot more planning, just a small bit of thought as to how it's going to hold itself up.  In this example, a z-level below should probably have supports in the same places, but actually it isn't strictly required.

[NW_Kohaku]

Well, what's the difference between a support and a wall? (Aside from that supports are more easily destroyed.)

Why wouldn't a person just leave a wall there that could be engraved, rather than a support?

[Jeoshua]

Point.  It would probably be more efficient to just leave the wall there.

Supports could be build out of any stone, metal, wood, etc however, so they could be used in this circumstance to have metal posts providing most of the support.

As to why not a constructed wall?  Why not indeed.  It's basically the same concept.

[NW_Kohaku]

Aww, that's not as fun.

Well, I came up with a couple of ideas, anyway.

1. You can walk through support tiles.  This could be useful in cases like digging through sand, where we might require supports in every tile before the roof caves in on you. 

2. Supports can weigh less (but also support less) than regular walls.  Especially if you could construct these things so that it weighs 1/5th as much as a wall, but supports 3/5ths as much weight, supports will have some justification. 

Incidentally, this could also be used in justifying some building materials like brick, concrete, and reinforced concrete. 

[Jeoshua]

The idea where supports weigh less does make sense.  Especially when you consider they can be made with blocks, which presumably have 1/4th the actual rock material as a boulder would (since you can make them 4:1 from a boulder).  If it's all being calculated based on weight vs yield, that makes supports ideal building material.

Doubly so if you can walk through them (but for the sake of argument, not place anything on them.  They're full tile height supports, not tables)


Concrete has been known since the romans, and I believe there have been some other civilizations that used similar stuff since the time of antiquity.  So it definitely fits into the whole scheme.  Reinforced concrete, on the other hand, not so much.  I could however see Adamantine supports being flooded over with magma and water, creating "Reinforced Obsidian" tho...

[NW_Kohaku]

Additives

Concrete additives have been used since Roman and Egyptian times, when it was discovered that adding volcanic ash to the mix allowed it to set under water. Similarly, the Romans knew that adding horse hair made concrete less liable to crack while it hardened and adding blood made it more frost-resistant.[8]

In modern times, researchers have experimented with the addition of other materials to create concrete with improved properties, such as higher strength or electrical conductivity

Steel rebar wouldn't have been used by the Romans mostly because steel was too precious for them, but it's not like "throw stuff in concrete" is somehow beyond dwarven comprehension. 

Of course, reinforced concrete is there for tensile strength, rather than compressive strength. 

I'm still thinking, however, that the game should be more worried about "this stone is jutting out too far" or "this ceiling arches too long without a support" more than "this wall bears too heavy a load, make a thicker wall".  The last of these is going to be something that may burden players a little much. 

I would, however, like to see the notion that you should build your walls on top of other walls in lower floors enforced.  If it's as simple as "build walls on top of other walls" without the notion that walls will be crushed from bearing too much weight without needing to make a foundation that is 7 tiles thick, then it's something I think people will be able to understand and play through without too much complaint. 

[Jeoshua]

Those examples do all come from about the same standpoint, tho.

If a stone is jutting out too far, it's because there isn't anything holding up it's weight.  And if a ceiling archjes too long without a support, in reality it's because there is too much weight not being distributed to the ground.

It really boils down to which direction you're looking at the problem from.  The Load Bearing method looks at the problem from a computing and programming point of view.  It's easier to say how to implement it because it models the whole thing from a very straightforward method.  On the other hand, the Number of Tiles Before Collapse method goes from the player's viewpoint.  It's not as easy to make a simple set of rules for, but much easier to communicate to people.

Ultimately, both systems should work in the same way.  The biggest difference is that while either method can model the system adqeuately if implemented correctly, only the Load Bearing method really takes into account higher and lower z-levels with really simple mechanics.

The system I outlined is, of course, not perfect.  I also thought it up in about 10 minutes and have no background in mechanical systems or architecture, and have yet to think of a situation that it really breaks down at.  So good enough for a quick thought, neh?

Either way, while the "X Tiles To Support" method worked well for the 2d game, here in the 3d world it really needs some kind of system that natively takes into account the z-levels above and below, and the possibilities that introduces.


One thing that I have come to determine that the "Cascade" system is missing is support from above.  A hard surface, like a cavern ceiling, needs to be able to be supported from above, suspended if you will.  Otherwise all caves will fall in immediately without a column support.

Only natural walls should be like this, however.  Constructed walls I just can't imagine being suspended like that.

[NW_Kohaku]

Chandeliers.

They hang like a stalagmite from a ceiling that supports them. 


Anyway, my point is that we would be looking at shear stress (how much weight can you put on a suspended beam before it snaps) and being strict about that, while being much more lenient on compressive stress (how much weight can you put on a brick before the brick crumbles to dust). 

Shear stress failures cause cave-ins when you have a bridge or ledge that is unsupported, and it falls.  Compressive stress failures cause the whole wall to be reduced to... I don't know, rubble or something, and crumble away while the whole building cascades down after it.  It's like how, when the World Trade Center's floors started to collapse, they fell onto the next floor down, which then also collapsed, which fell to the next floor down, which also collapsed, etc, etc.

Shear stress causes things that are overhanging to fall down, but compressive stresses cause the whole damn mountain to be reduced to rubble. 

[Jeoshua]

I'm not touching the WTC event with a 10 foot pole.  But I get your point, and it's a good one.

What if both forces were taken into account?  COMPRESSIVE strengths for the downward load that walls and supports could offer, and TENSILE stengths for the horizontal support like floors and ceilings?  If both were taken into account, and both were able to keep a block from falling, then there would be many ways to support the different structures.

It does eliminate the simplicity of a Weight Cascade like I outlined, but it does end up much more robust.

I think the sweet spot in any idea about this would be it's Dwarfiness.  It should be easy enough to explain the gist of it, but complex enough that people could spend a lot of time discussing it's finer points... like anything that's currently in Dwarf Fortress.

[NW_Kohaku]

I'm fairly sure what we want is the shear fracture point.  That's the "how much you need to bend it before it snaps" part.

Tensile strength would be of more importance if we are modeling the wind force upon a building.  As cool as it would be to recreate Galloping Gertie, I think that would require far too much simulation on the part of the game, and would only annoy most people. 

There would be shear yield, which is where the structure starts deforming, and then it would gradually hit the shear fracture point, where the cave-in occurs. 

Shear yield wouldn't be tested by walls per se, it's tested by floors or other things overhanging.  So if you had this sort of wall with an overhanging floor/walls built out, viewed from the side:

#
####
#
#####

Then the thing being tested the most for shear yield or shear fracture is the underlined # (although the other # pieces built to its right would also have shear stress, the one that would snap first is the underlined).  The actual wall part isn't being tested for shear stress, it would be tested for compressive stress, or not tested at all.

[Jeoshua]

Ah yes.  Shear Forces.  Where is my head? Tensile is for things like rope.  Shear is for breakage.

And yes, swinging rope bridges are far outside the scope of any cave-in mechanics.  It would be interesting to model but it's an entirely separate thing.  We're trying to think of things for architecture, not true-physics simulationism here.

The question is, what governs the chandeliers and other hanging natural stones?  Some things that are currently randomly generated would probably fall as it is now... and probably should realistically.  But they should nonetheless be somewhat possible, if a bit more dangerous for someone to try and construct.

So far the forces we're considering are:

• COMPRESSIVE forces - For vertical support (Columns and Load Bearing walls)
• SHEAR forces - For lateral support (Floors, ceilings, and overhands)

There are also:
MOLAR_MASS
IMPACT_YIELD, IMPACT_FRACTURE, and IMPACT_STRAIN_AT_YIELD
TENSILE_YIELD, TENSILE_FRACTURE, and TENSILE_STRAIN_AT_YIELD
TORSION_YIELD, TORSION_FRACTURE, and TORSION_STRAIN_AT_YIELD
BENDING_YIELD, BENDING_FRACTURE, and BENDING_STRAIN_AT_YIELD

Some of these could possibly find their way into a calculation.  It does make the whole simulation less predictable, but ultimately more realistic, to find a realistic way to use them.

[NW_Kohaku]

Shear is when you put a load on something, and how much weight it takes for a rod to bend or break when you put weight on a rod supported on both ends.

Compressive is how much it resists going "squish" under a lot of weight when you load things on top of it when it is supported underneath (or when you crush it between two things)

Tensile strength is how much something will "spring" back into shape after being bent.  The difference between tensile strength and shear yield strength is in whether the object "elastically" deforms (goes back to its original shape) or "plastically" deforms (stays bent). 

Torsion is if you twist something instead of just bend it.

The difference between yield and fracture is that fracture is the point where the thing just plain fails and snaps or is crushed, while yield is where it is merely deformed.  However, as something deforms, the resistance it has to further deformation becomes weaker, and it gets closer to simply breaking entirely. 

Molar mass is basically related to density on the molecular level.  It is the mass per mole (a specific number of molecules), and so it's a function of the atomic weights of the constituent atoms of the molecules of a material.  It's something that all these shear strengths and densities and such derive from in real life, but it's a distant thing in architecture - we only need the derived numbers.


Basically, I think all we really need is shear fracture, with maybe some compressive fracture, but in a way that doesn't become too imposing on the player. 

Shear fracture would be checked on how the "ledge" that juts out is supported (or not) when it reaches over.  Making an arch, where you have a thicker wall that supports less weight the further away from the load-bearing wall you go, would be a good way to support weights over large distances without supports. 

The "chandelier" idea would be handled by this - you just add its weight onto the supporting ceilings for tests of shear forces on how well the ceilings hold up. 

Doing this, we wouldn't be testing if the load-bearing walls are going to be crushed at all, just testing if the "ceilings" or "floors" have the structural integrity to not fail as they are extended over unsupported space.


If you absolutely must do load-bearing walls where those walls can crumble under the shear weight of the walls they are supporting above them, then it's a matter of dividing out that mass from above across all supporting tiles below.  That's where you get the "cascade" calculations, where stress flows downwards from the top, with each tile adding its own mass onto the stress that flows downward, but where it gets divided out across other stones if there are multiple tiles of support. 

Presumably, tiles that have other stone walls beside it should never fail, no matter how much pressure they are under.  They might be under enough pressure to deform, but if there's no place to deform to, they just get compressed.  This is pretty much the point where you put enough pressure on something to turn sand into sandstone or coal into diamonds or something, but we don't really need to model that.

[King Mir]

The problem is, again, what about people who want to plan out a large castle or tower, or are in the middle of excavating a huge dining hall. 

If you just start designating without paying attention to it, you could collapse your whole fort in on itself. 

If you start building a tower, you have no in-game clue as to how strong or how heavy any given building material is. 

Before you can demand that players start having "realistic" thicker foundations and use lighter building materials at the tops of their towers, you have to actually tell players which materials are stronger, somehow.

Otherwise, how am I supposed to know if dolomite is a better wall material than limestone or granite?

The same way you know that iron is about as strong as bronze, and that copper armor is inferior to all other metal, but much better than leather or wood.

[Jeoshua]

Except much easier, since nearly all stone uses STONE_TEMPLATE for almost all of it's actual mechanical stress values, all soil uses SOIL_TEMPLATE which is nearly identical to STONE_TEMPLATE, and all metals use METAL_TEMPLATE for everything we've been discussing as useful in these kind of calculations.

So it's basically METAL > STONE = SOIL > WOOD

[King Mir]

Wood should be better than soil for building things out of.