Basic damage modeling for ships

[twistking]

Of course this would also apply to modern DCS, but especially the PTO will rise or fall with ship damage modelling.

I think we can all agree, that a purely hitpoint based system simply won't cut it.

I think ship modelling should at least have the following features:

- Damage must be seperated from flooding (buoyancy), meaning a "destroyed" ship can still float and a "not destroyed" ship can still flood and sink
- Flooding must take compartments into account - at least in a simplyfied manner
- Different parts of the ships take damage and fail individually
- Ideally there was also some limited (!) self-repair (stop flooding, stop fires) tied to a "crew status" value. "Crew status" would decrease through hits, with lower values limiting and finally stopping self-repair capability of the ship. If "crew status" is completely depleted, the ship is abandoned.
- With this there would be three basic kill types: 1. Sinking, because of flooding a set amount of compartments. 2. Ship fuctionally destroyed (superstructure, weapons etc) but still floating without power. 3. Ship abandoned because of depleted "crew status". Similar to previous point, but visibly different (less destruction) and maybe a more realistic result for a bigger, heavily armored ship (in reality high casulties and spreading fire could trigger an evacuation without the ship being completely destroyed or sunk).

 

Edited January 7, 2024 by twistking

[AG-51_Razor]

"Damage must be seperated from flooding (buoyancy), meaning a "destroyed" ship can still float and a "not destroyed" ship can still flood and sink"

I get the first part, about a "destroyed" ship could still remain afloat, but the second part about a "not destroyed" ship can still flood and sink doesn't make any sense to me. I recon that the whole point of attacking a ship is to sink her, thus "destroying" her.

That said, I do agree with your comments over all. ED could definitely use some improvements in the shipping department as far as damage is concerned.

[Silver_Dragon]

Not only, Fire damage need take account with the other problem of magazine explosions. Fire and Flodding need take account to not destroy by control damage teams on a ship before the "damage" reach to zero.

[Gunfreak]

48 minutes ago, AG-51_Razor said:

"Damage must be seperated from flooding (buoyancy), meaning a "destroyed" ship can still float and a "not destroyed" ship can still flood and sink"

I get the first part, about a "destroyed" ship could still remain afloat, but the second part about a "not destroyed" ship can still flood and sink doesn't make any sense to me. I recon that the whole point of attacking a ship is to sink her, thus "destroying" her.

That said, I do agree with your comments over all. ED could definitely use some improvements in the shipping department as far as damage is concerned.

I think he means it doesn't explode, as in torpedo hole in the bow can sink a ship without (blowing it up)

[twistking]

2 hours ago, AG-51_Razor said:

"Damage must be seperated from flooding (buoyancy), meaning a "destroyed" ship can still float and a "not destroyed" ship can still flood and sink"

I get the first part, about a "destroyed" ship could still remain afloat, but the second part about a "not destroyed" ship can still flood and sink doesn't make any sense to me. I recon that the whole point of attacking a ship is to sink her, thus "destroying" her.

That said, I do agree with your comments over all. ED could definitely use some improvements in the shipping department as far as damage is concerned.

1 hour ago, Gunfreak said:

I think he means it doesn't explode, as in torpedo hole in the bow can sink a ship without (blowing it up)

Yes, basically what Gunfreak said. I could have worded it better. Of course, once a ship is sunk, it can be considered destroyed. In a simple hitpoint-based system sinking would simply be an animation step at 0 HP (the ship completely battered, everything burning etc). However a lucky hit (or two) should be able to "mortally" wound a ship even when it's superstucture is still in pristine condition.

2 hours ago, Silver_Dragon said:

Not only, Fire damage need take account with the other problem of magazine explosions. Fire and Flodding need take account to not destroy by control damage teams on a ship before the "damage" reach to zero.

Yes, magazine explosions would also be a good feature.

[Gunfreak]

I don't think we can expect too much.(like compartments and damage control)

Not sure we can expect anything really.

But of they wanted some simulations.

I'd first add 

1. Hull damage that can sink a ship

2. Weapons system can be individually damaged (this is already on some ship mods)

3. Carrier deck damage (this is already on carriers to some degree )

4. Possibility of catastrophic damage (basically if bomb hits magazines, which is a designated spot where the magazines were on that particular ship type, then boom boom bye bye for ship)

Even this is probably way more detailed then anything I expect to see in DCS for many many years (if ever)

[twistking]

12 minutes ago, Gunfreak said:

1. Hull damage that can sink a ship

2. Weapons system can be individually damaged (this is already on some ship mods)

3. Carrier deck damage (this is already on carriers to some degree )

4. Possibility of catastrophic damage (basically if bomb hits magazines, which is a designated spot where the magazines were on that particular ship type, then boom boom bye bye for ship)

That would be a good MVP indeed. Flooding tied to a Hull HP value, will however mean that ships will always sink in the same attitude. Compartment flooding would allow ships to sink in a rudimentary "simulated" way, maybe even roll over.

[WinterH]

As a steam&steel era ship nerd, I'll still say no, hit points should be quite enough (with caveats). Chasing perfection is often counterproductive for a good result, we need to remember DCS is foremost a flight sim, and a rather resource hungry one at that.

Modeling compartments, calculating damage and buoyancy separately for each compartment/component of each ship in a busy mission, is overkill in my opinion.

Now, I do agree some improvement is needed, hence the caveats. Maybe only track damage for semi-arbitrarily defined compartments (as opposed to perfectly blueprint fitting ones), and consider ship killed if x number of compartments are dead should be something for example. Also, definitely individual equipment above the deck should be destructible. Even if simplified by a "zone with hit points" rather than individual gun/sensor's detailed damage model, we'd need to be able to disable AA guns, sensors, fire directors (with hopefully expected effect on the accuracy) etc, as opposed to current "damage anywhere on the ship enough to maybe disable weapons" deal.

But a detailed damage model for each ship and its sub modules considering physical/fire/flooding damages is both overkill for the purpose, and not called for in context of DCS in my opinion.

Even fairly realistic naval warfare games do a lot of approximation and abstraction modeling the survivability of ships, and those are almost the only things they need to model, or at least one of the main things anyway. DCS will need to cope with (hopefully much improved) AI of many ground/naval/air units, complex flight and systems modeling of many user flown aircraft, and the high polished graphical representation of both those things and the world they occupy. A naval sin level damage modeling for ships is overkill and would compete with these I'd think. And self-repair, damage control etc is definitely overkill in my opinion.

Don't get me wrong, I do definitely agree that some improvement over the current state would be very welcome, and frankly somewhat necessary. Just not to the extreme you suggest twistking, not in my opinion anyway.

[greco.bernardi]

A new advanced damage model  for ships will be a must important adiction for a proper recreation of navy battles like a carrier battles.
Will be nice to have crew managing the AAA weapons on ships too and the possibility to disable some weapons or system.

A ship damaged by a torpedo atack heeling because of some compartiment flood... trying too keep in flight will be wounderfull.

And will be nice to have AI ships turning to avoid torpedo or dive bombers.

Edited January 7, 2024 by greco.bernardi

[twistking]

18 hours ago, WinterH said:

As a steam&steel era ship nerd, I'll still say no, hit points should be quite enough (with caveats). Chasing perfection is often counterproductive for a good result, we need to remember DCS is foremost a flight sim, and a rather resource hungry one at that.

Modeling compartments, calculating damage and buoyancy separately for each compartment/component of each ship in a busy mission, is overkill in my opinion.

Now, I do agree some improvement is needed, hence the caveats. Maybe only track damage for semi-arbitrarily defined compartments (as opposed to perfectly blueprint fitting ones), and consider ship killed if x number of compartments are dead should be something for example. Also, definitely individual equipment above the deck should be destructible. Even if simplified by a "zone with hit points" rather than individual gun/sensor's detailed damage model, we'd need to be able to disable AA guns, sensors, fire directors (with hopefully expected effect on the accuracy) etc, as opposed to current "damage anywhere on the ship enough to maybe disable weapons" deal.

But a detailed damage model for each ship and its sub modules considering physical/fire/flooding damages is both overkill for the purpose, and not called for in context of DCS in my opinion.

Even fairly realistic naval warfare games do a lot of approximation and abstraction modeling the survivability of ships, and those are almost the only things they need to model, or at least one of the main things anyway. DCS will need to cope with (hopefully much improved) AI of many ground/naval/air units, complex flight and systems modeling of many user flown aircraft, and the high polished graphical representation of both those things and the world they occupy. A naval sin level damage modeling for ships is overkill and would compete with these I'd think. And self-repair, damage control etc is definitely overkill in my opinion.

Don't get me wrong, I do definitely agree that some improvement over the current state would be very welcome, and frankly somewhat necessary. Just not to the extreme you suggest twistking, not in my opinion anyway.

I think you're reading more complexity into my suggestion, than i ment. Everything i proposed can be hugely abstracted and still work "good enough" for a flightsim. Buoyancy does not need to be physically simulated. It could be a look-up table or other rudimentary system. Damage to superstructure and weapons could be done by dividing the ships in a few logical parts and having basic hitpoints for each part, assosciated with visible damage to the part and loss of functionality of systems in that area. Magazine explosion could be a dice roll with some simple condition checks for penetration.

It's not about having a true-to-life simulation from the ships perspective, but having the simulation complex enough for the pilots in the air to make weapon selection and tactics a menaingfull decision. F.e.: It should matter if i deliver a certain amount of HE with a salvo of rockets, or if i deliver the same amount with one big bombs with a delayed fuze. A torpedo should be a different beast again and not just another way to deplete hitpoints.

Such a system is a bit of work to develop surely, but it's not computational expensive.

Edited January 8, 2024 by twistking

[Northstar98]

Yeah, hard agree - the current system is very lacklustre, it doesn't even approach the fidelity of the pre-WWII era damage model, which is already problematic for aircraft.

 

With that said, brace yourselves everybody, humungous wall of text incoming!

On 1/7/2024 at 8:21 PM, WinterH said:

Modeling compartments, calculating damage and buoyancy separately for each compartment/component of each ship in a busy mission, is overkill in my opinion.

Calculating damage could still be done with a hitpoint based system. As for buoyancy, I doubt anyone here have something that is any higher fidelity than the simplified flight model (i.e. only consider basic, abstracted forces) in mind.

Buoyancy is pretty easy here because it's purely arithmetic so long as we know the volume of water being displaced, which we do because we know the displacement of ships and their draught. All we need now is some abstracted compartment system, which could be some number of cuboid (to keep calculations trivial) compartments; the weight and buoyancy of the vessel can be uniformally distributed between them and forces can be treated as acting in the centre of each compartment.

How much a compartment is flooded could also easily work with a points based system - each compartment has its own number of hit points, how many it loses determines the rate of flooding. The compartment system described above makes things fairly easy for determining weight and buoyancy (so long as we know the density of water and the volume of the compartments - both of which we do).

There's no reason why fire couldn't work in the same way. The only thing with fire is that it should progressively damage internal components and the compartment itself (both hitpoints based, so shouldn't be particularly taxing).

As for damage control, all that would do is remove "sinking" and "fire damage" points at some rate. What fires should do is damage any internal components (which again, they themselves would have their own hitpoints based system, which is how components already work).

For superstructure compartments, flooding is out of the equation and you're left with something similar to what we have now with some aircraft + fire damage.

 

All of these are fairly simple arithmetic operations from a modelling perspective - I doubt it's any higher fidelity than simplified flight models for most AI aircraft. But it should be perfectly adequate for our purposes.

On 1/7/2024 at 8:21 PM, WinterH said:

and consider ship killed if x number of compartments are dead should be something for example.

Well, hopefully DCS would be able to track mission-kills separated from catastrophic ones - it's already problematic for ground vehicles.

On 1/7/2024 at 8:21 PM, WinterH said:

Even if simplified by a "zone with hit points" rather than individual gun/sensor's detailed damage model, we'd need to be able to disable AA guns, sensors, fire directors (with hopefully expected effect on the accuracy) etc, as opposed to current "damage anywhere on the ship enough to maybe disable weapons" deal.

Some ships already have some external components able to be damaged (usually guns/launchers and at least one class of fire-control radar), but it's quite inconsistent. A hitpoints based system is perfectly adequate so long as it's implemented properly. As for states, it only needs to keep track of operational (no damage) disabled and destroyed (hitpoints fully depleted), with the consequences of damage to whatever component considered.

But it should also apply to internal components - at the very least command and control (bridge & CIC/operations room), engines and magazines.

But as for "detailed damage model", I don't think anybody is expecting anything like the WWII damage model, where even individual engine cylinders are modelled. Obviously from a fidelity standpoint the more the merrier, but this could quite easily be simplified to just include the essentials.

On 1/7/2024 at 8:21 PM, WinterH said:

Even fairly realistic naval warfare games do a lot of approximation and abstraction modeling the survivability of ships

And a lot of approximation and abstraction would be perfectly serviceable here.

On 1/7/2024 at 8:21 PM, WinterH said:

DCS will need to cope with (hopefully much improved) AI of many ground/naval/air units

It will, but this shouldn't be computationally expensive compared to those - we already have a far higher fidelity damage model than what I've laid out in the WWII damage model and the simplified flight and pre-WWII damage model, at least in principle and fidelity, already does what such a system would need (i.e. consider basic forces, how large they are and where they act).

On 1/7/2024 at 8:21 PM, WinterH said:

complex flight and systems modeling of many user flown aircraft

Why would it ever need to cope with more than one aircraft? I thought in multiplayer, all calculations relating to an aircraft's flight and systems modelling is done client-side? I'm aware that some things fall on the host's machine but AFAIK the flight models of other player aircraft aren't one of them, as it should.

On 1/7/2024 at 8:21 PM, WinterH said:

and the high polished graphical representation of both those things and the world they occupy.

This is more problematic but DCS already should be employing methods that deal with this - things like LODs for instance.

On 1/7/2024 at 8:21 PM, WinterH said:

Don't get me wrong, I do definitely agree that some improvement over the current state would be very welcome, and frankly somewhat necessary. Just not to the extreme you suggest twistking, not in my opinion anyway.

Well, I personally don't see what the compromise would be between what we have now and the improvements proposed beyond just making the current system more consistent, which IMO, while definitely desireable as consistency isn't something I'd praise DCS for, wouldn't really lead to tangible improvements.

Edited January 11, 2024 by Northstar98
grammar, minor addendums

[draconus]

Damage modeling is needed for all ships, not only PTO/WW2. We can start with something simple, then add the complexity, but the hitboxes have to be done as complex as possible from the start - no need to go back and rework it again and again. I wouldn't worry about performance apart from more memory requirements (damaged parts need 3D models and textures). We have MT and calculations doesn't have to be done every frame - 1 second is a very long time for our CPUs and ship sinking is usually a very long process, also they don't go down en masse. Fully fledged DM for a ship is a huge development task though.

[twistking]

13 hours ago, draconus said:

Damage modeling is needed for all ships, not only PTO/WW2. We can start with something simple, then add the complexity, but the hitboxes have to be done as complex as possible from the start - no need to go back and rework it again and again. I wouldn't worry about performance apart from more memory requirements (damaged parts need 3D models and textures). We have MT and calculations doesn't have to be done every frame - 1 second is a very long time for our CPUs and ship sinking is usually a very long process, also they don't go down en masse. Fully fledged DM for a ship is a huge development task though.

I agree. And i can't stress enough that computationally it would be a non-issue, if done correctly. Development time is the cost here - not frame time.

[rkk01]

Compartments / flooding could be simplified along the lines of focstle, midships, aft - to allow bow or stern first sinking,  port / starboard for list / capsize.

Perhaps separate damage hit boxes for magazines, machinery spaces and CV avgas bunkers

I guess hull and deck playing would need an “Armour” value that would repel or be defeated by set ordnance. Below waterline hits should be able to cause flooding damage

 

[peachmonkey]

if you guys are going down this rabbit hole don't forget how they sunk Tirpitz in the fjords. There were direct hits, yes, but also quite a few that detonated close to the bow of the ship as well impacting the integrity/strength of the outside shell, which all of it contributed to Tirpitz capsizing.

Talking about splash damage IN the water.

[Silver_Dragon]

4 minutes ago, peachmonkey said:

if you guys are going down this rabbit hole don't forget how they sunk Tirpitz in the fjords. There were direct hits, yes, but also quite a few that detonated close to the bow of the ship as well impacting the integrity/strength of the outside shell, which all of it contributed to Tirpitz capsizing.

Talking about splash damage IN the water.

Underwater explosions has required to Depth charges, bombs, shells, and torpedoes.

[upyr1]

On 1/6/2024 at 9:01 PM, twistking said:

Of course this would also apply to modern DCS, but especially the PTO will rise or fall with ship damage modelling.

I think we can all agree, that a purely hitpoint based system simply won't cut it.

Agreed, we've needed a massive overhaul of DCS naval units for a while especially with the Marianas. I think a damage model over haul would be part of the necessary prep work for possible naval modules. 

On 1/6/2024 at 9:01 PM, twistking said:

I think ship modelling should at least have the following features:

- Damage must be seperated from flooding (buoyancy), meaning a "destroyed" ship can still float and a "not destroyed" ship can still flood and sink

Battleship Row is a good example of the second. Except for ARIZONA and OKLAHOMA all of the active battleships that got sunk were returned to service. AZ blew up after a bomb penetrated her armor and hit a magazine and OK capsized. 

On 1/6/2024 at 9:01 PM, twistking said:

- Flooding must take compartments into account - at least in a simplyfied manner
- Different parts of the ships take damage and fail individually

The ideal situations not only would we have damage models for differnt part of the ships, but we would also have things like armor factored in as well.  For example an attack against a battleship might have 0 chance of penetrating the armor while it might be able to knock out an AA gun or a radar dish.  Then ideally I would also like to see internal systems damage as well, for example one of the reasons HMS  SHEFFIELD was lost was because her fire main was taken out and I speculate this may have happened with MOSKVA as well.  

On 1/6/2024 at 9:01 PM, twistking said:

- Ideally there was also some limited (!) self-repair (stop flooding, stop fires) tied to a "crew status" value. "Crew status" would decrease through hits, with lower values limiting and finally stopping self-repair capability of the ship. If "crew status" is completely depleted, the ship is abandoned.
- With this there would be three basic kill types: 1. Sinking, because of flooding a set amount of compartments. 2. Ship fuctionally destroyed (superstructure, weapons etc) but still floating without power. 3. Ship abandoned because of depleted "crew status". Similar to previous point, but visibly different (less destruction) and maybe a more realistic result for a bigger, heavily armored ship (in reality high casulties and spreading fire could trigger an evacuation without the ship being completely destroyed or sunk).

Crew skill would definitely be an issue here, STARK and the 1980s SAMUEL B. ROBERTS are good examples here, as crew skill was part of the reason they aren't at the bottom of the Persian Gulf.  I figure an expert crew might have a lumbar yard on board like NEW JERSEY, they are still finding damage control lockers on her and she's been a museum for 20 years