:-) this is graph of sonar self-noise, so it's conected with hydrodynamic, "flow" noise. Well, it's even not flow noise itself but "sonar self noise" ! Flow noise at zero speed should be rather zero don't you think ? So I guess on your graph it's combination of sonar self noise with flow noise.
First, nobody said flow noise is linear :-). But the noise generated by a sub so combination of few kinds of noise, al low speeds machinery noise predominates, at high speeds in modern subs rather flow noise predominates. The flow noise is not linear. But this graph don't look good for me. Because from what I know the flow noise in general is proportional to sixth power of the speed (may be other power at lower speeds maybe). So it's raising curve as you think. BUT it is not raising curve if you plot it on a logarythmic graph !! It becomes the opposite then, because logarytm or a^x is stronger function than even x^6. On your graph the noise is in dB and speed in kts - and the curve is raising. This would lead to an absurd on the very right side of the graph, because more you are going to the right (higher speed) faster the noise (in dB) raises, so at some point increase of speed by 1kts would lead to several dB of noise increase so noise would double or raise 10 times with 1kts of speed increase ! Don't you feel something is wrong here ?
The graph of flow noise can't raise higher and higher on the right.
So even if this graph of sonar self noise is correct (because for example it's determined not only by flow noise but other things) it has not much to do with submarine self-noise which is combination of flow noise and few other kinds of noise. If sonar self noise is combination of constant base 50dB + flow noise, then it could look like on your graph but you'd be surprised how it looks more to the right :-).
Well ok I have found my Excel spreadsheet with noise calculations and made a combination of constant base NL and a flow noise. After a while I have found a combination of parameters that give quite similar result, look (click on the lower graph to enlarge):

http://www.fas.org/man/dod-101/navy/...s/IMG00006.GIF

http://img373.imageshack.us/img373/2...tant2ou.th.jpg

Look at the red line, looks quite similar. But the first impression is wrong, when you take a second look at what happens more to the right and understand what's happening - it's actually a dB scale sum of constant 20dB noise with a flow noise. Look at second graph, it's same function but showed more to the right and with the flow noise part showed in black, as you see the red line is starting from base 20dB but later it becomes more and more similar to the black line.

http://img373.imageshack.us/img373/2...ant28gx.th.jpg

The black line is flow noise and I guess it's quite different from what you could expect. But the red line, sum of 20dB and flow noise at first look quite different than it becomes later. So not always looking at begin of an graph can give you good understanding of what will happen later, if you don't know what function or sum of functions the graph shows. The begin may look quite different from the rest of graph.

But I think the graph you showed is not very precise and maybe also the sonar flow noise at low speeds is little different function because it don't look like speed^6 function. So it's either more complicated (and even less connected with submarine noise) or not very precise...

sure Amizaur : my graph wasn't very accurate, it was just an example.
But personnal experience abord sailing ships and my love for physics (even if I'm not a physicina myself) just show me there is real gaps in this matter
For example, on my ship, well ... my father ship, a catana of 44ft, I could see it's quite easy to go to 10 knts, even with low wind speed (15 or 17 knts), without near any perturbations at the tail of the ship, and the perturbations will start around 11 to 12 knts.
From that speed, we really need more wind to get over this gap (our personnal best speed is 21.3knts on this boat, with 40/45 knts of wind "au portant" and the smallest sail)
It's simply due to the hull shape.

I'm convinced there is some gap also for sub hulls.
Because they are also subject to physical laws.
So, gap changing drastically the noise generated.
There is a before and after the gap
after the gap, water perturbations start and make much more noise than the noise generated by the sub itself.
This gap is not modelised at this time
and this is domageable for tactics.
because I'm convinced a good commander of a real sub KNOW this gap and know when he can go over this or when he can't, because he will risk counter-detection

This gap should be VERY different, depending on sub tech
Well all read Akula is as quiet as US subs AT LOW SPEED and NOT on high speed.
I think it's just because this gap is reached earlier with the Akula than with a 688i and furthermore with the Seawolf.

Seawolf is known to be "silent until 20 knts"
20knts is his gap, before he is really stealthy, after he make much more noise.
I think this gap should be around 12 or 14 knts for 688i, and beetween 8 to 10 knts with Akula.
At this time this isn't modelised at all
We have a progressive and linear curve that doesnt' reflect this kind of real problems.

Of course, these data should be one of the most secret ones.
But with some talk and tought, we could find something much closer to the real thing than what we have actually.
I'm still convinced the original data were closer to the real thing than what we have from the sound vs speed fix ... or with the LWAMI.

I say you again that original data was ZERO, not any noise increase with speed at all :-D. The SCS noise-speed fix changed this to 10 points (so 20dB) of linear function of speed for SSNs.
I'm not sure about what "gap" you are talking here. If you have water drag in mind (which is quite complicated function of the speed in the water) or difference between laminar and turbulent flow maybe ?
The noise generated is related to both drag and flow kind (turbulent being more noisy) but don't know if it's related so directly to make same plot and to make assumptions about noise created by an object in water from the knowledge of it's drag characteristics. And even if you have plotted function of object's drag in the water vs speed on logarythmic scale for drag, you'd be surprised how the graph looks :-). It bends in the other side than you expect :-).
For sure the Seawolf (and every modern sub) is designed to reduce turbulence and reduce flow noise (or rather to increase the speed at which the flow noise becomes noticeable). And if the machinery noise is low and not increase much with speed, and the flow noise is reduced so that it becomes to show up only at 20kts, then you get a quieter sub than can run faster being quiet. But this would also mean being quieter from competition at lower speeds :-). The situation you describe would occur if there is a minimum amount of machinery noise that CAN'T be reduced even at stop (why?), but on the other hand it is dampened so good that it don't increase with speed almost at all and stays on this lowest level up to 20kts. And then at 20kts the flow noise shows up and raises further with the speed. So base noise level at 0, 5, 10, 15 kts, and then flow noise becomes to show and kicks in at 20kts and higher. Well little strange, if there is some machinery that makes some noise at idle (and you can't reduce it to non-measurable value) then the machinery would make increased noise at half speed or full speed... Or do we assume that every piece of Seawolf's propulsion (reactor and it's cooling, steam generators and turbines, reduction gear and shaft, propulsor) is working or rotating with same speed at 5kts than when at 25kts ?
What is true that the machinery noise raises slower than flow noise and that's why the flow noise predominate at some speed. But I don't believe that it's 110db, 110db, 110db and suddenly starts to raise 120 130 140. It may raise slower at low speeds, and become to raise faster when flow noise predominates. It may be at 20kts in Seawolf and 10kts with 688I maybe, and maybe not. We don't know at what speed the flow noise predominates for modern subs. But I have assummed that it's about 20kts for Seawolf and about 10kts for 688I and set the machinery and flow noise graphs to reflect this. Currently I can't model this in speed-noise curves (because they ARE linear in DW from the very start, it was SC were they were non-linear) but I know it CAN be done, because the FFG has non-linear curve, or more precisely have some flat area on the beginning of the later-linear graph. I can't make sub curves same way because this affects sub minimum speed. It can be compensated and it is compensated for FFG (it can stop) but I don't know exactly how and can't do this. But I plan to find how it was done for FFG and then I could add flat area on beginning of sub's noise graph, even 15kts of it :-). I will plot supposed sub noise profile and then find what possible to achieve DW noise curve would fit best to it. Possible to achieve in DW are linear with some flat area on beginning. Non linear noise function was in Sub Command only and was not very fortunate because even flow noise curve raises sharply at some speed range, but then becomes to flatten in dB scale. In SC it raised faster and faster with speed hitting max already before flank speed.
I think current model is better and more accurate, if I knew how to compensate the minimum RPM value effect on minimum speed it would be even better with flat beginning possible.

I didn't find again the post when I tested this (dated, as my screens, from mars of this year)
but still have the screens that prove there WAS change depending on the speed =>

With Stock DW
KILO at 12 knts

http://okof4.free.fr/images/DW/screenshots/KILO_12.jpg

KILO at 17 knts

http://okof4.free.fr/images/DW/screenshots/KILO_17.jpg

you can see a real difference here.
So, it wasn't 0 Amizaur ...
As I tested it, I remember there was a real huge gap at some speed with ALL subs, not only the KILO.

I think pictures speak by themselves.

it's a fact known for all nukes
even at complete stop they still make noise, and there is very few difference beetween a nuke at stop and a nuke at 5 or 7 knts

But I don't think you get my point :
As I told you, there is always, on every hull, a time where you start creating hydrodynamical perturbations, when the laminar flow start to become not so laminar, and here turbulence are created.
As I told you, I could see that each time I go sailing, on the rear of my boat.
that's because my hulls (its a multihull) are optimised for cruise at 10 to 11 knts.
If I go over this speed, even if I could go much faster (well ... its still a sailing boat anyway ...) I will have much more turbulence in my wakes.
but before 10 knts, if you look at the wakes, you have the impression that the ship is stopped
and suddenly, if you go over that speed, you see wakes coming.
It's the time the hydrodynamical flow stop to be absolutly laminar along my hulls.
Not only this slowly brake the increase of speed, but also from this point, I start to need more energy (from the wind) to increase the speed than I needed before 10 knts.
Just because I've got out of the optimal flow, starting turbulence.
And turbulence for subs is noise.

So, for subs, there is also a critical point like this.
A point where a real commander will know he will produce much more noise.
And this is an important tactical thing.
Tactical thing that doesn't exist at this time in DW, but existed in stock DW, as the screens above could show you

I have to say that anyway i don't know if stock DW had good values ...
I just saw that there was a gap, that's all.

So you have only those two pictures, not a test mission where it could be repeated and measured ? From this one case ? I'm sorry but I have done not one, but dozen of tests, in different scenarios and different ways to measure noise level and I'm absolutely sure it was constant in 1.01. There were also later available some ways to measure this directly, something like Debugoutput showing platform noise level, results were the same.
The difference could be because of cavitation, but you said it wasn't.
I see there are two pictures, how were they made ? Was it a Kilo that run at 12kts and then increased speed to 17kts ? Or was it the same mission run two times, once with Kilo set to 12kts, and second time with Kilo set to 17kts ? Because you can't compare results from 2 different runs even of the same scenario. Water conditions are different each time, layer depth is different and detection ranges can vary, so the strentgh of the signal.
If you can replicate the situation and send me the mission I could check it, at least in few ways. But that would have to be two Kilos in same mission, one going 12kts and the other 17kts and generating different amount of noise. Or Kilo going 12, and after reaching a waypoint speeding to 17kts but in the same mission.

About the noise becoming stronger with turbulent flow - of course but I'm not sure the difference is so sharp UNDERWATER, where is no air which can mix with water causing bubbles, splashing and noise. And the speed at which flow becomes turbulent can be different at differents part of the sub, just like local Mach number can be different around differents parts of a plane.
The flow probably becomes turbulent on low scale (close to submarine skin) much sooner than 20kts, only the shapes of hull are designed to keep the flow smooth and produce less noise. It's my guess of course but it's hard for me to imagine fully laminar water flow around sub surface at speed of 15kts or more. I think the noise level of flow would increase more or less smoothly, if the shape of sub is designed properly (so modern sub).
The really sharp increase in noise level underwater is connected with cavitation rather, which occurs only at much higher speeds, like for propeller blade or Shkval torpedo.
But as I said you could probably clearly see on the graph of broadband noise where the flow noise becomes to predominate other noises - the plot would become to raise faster from there.

I took this screens with a scenario I made 8 month ago
i used twice the same scenario, just changed the speed of the KILO.

Whether you like it or not, I got this result.
Now try something if you don't believe me =>
desinstall DW and reinstall it without patching
then proceed to this very simple verification : put a KILO at 1000 ft and do the test.
I did it with both KILO and 688i and had the same result, but I can't remember the speed at which the 688i started to make big noise, this test was made 8 month ago, BEFORE the sound fix.
Anyway, this is another subject, I don't care about older values, i care about actual ones, which are wrong.

About turbulence, I explain you what I COULD SEE each time i'm sailing, this mean beetween 4 and 6 times a year.
but maybe you have more experience than me on this subject, even if I sail for 35 years on many different ships.
Now, as I see you don't even trust me about my tests (at least you didn't said I lies ...) I don't see any reason to continue this discussion ...

but fact are :

1) ANY NAVAL HULL in the world, made by human (i don't know about ET ones ...) have a critical point at which turbulence start to be created, and this point depends DIRECTLY to the type of hull (the form and the length) and allow most mdern ship to have best performance (ie step forward) in this matter, because more research was made to delay this critical point.

2) at this critical point, the non linear flow not only produce turbulence but start also to reduce the hydrodynamical performance of the ship (mean it need more power than before to increase the speed), if you stll don't believe me, just read some physics basics.

3) when turbulence are created, not only more opposition to advancing is generated, but also NOISE is generated, and MUCH MORE NOISE than before turbulence occurs.
Simple example is the washout on sonar when you reach 14knts with the 688i, when it is 8 knts with the Akula
You didn't ask yourself why there is difference here ? it's mainly due to the hull hydrodynamical performance, not only the sonar sensitivity.

4) NOTHING OF THIS is modelised in actual DW nor LWAMI, and didn't even approach this, all sound profiles are absolutly linear actually, that is not correct and not realistic.

5) I never said original was realistic, but i said It was closer to the real thing than what we have actually, because i could saw a gap I don't see anymore in the actual DW with sound fix


You see, military are actually trying to reproduce the skin of the shark to use it for subs skins, and on competition sailing boats, and you know why ?
guess ...
to delay to the maximum the time the turbulence start to be generated.
because when turbulence start to be generated, this degrade the hydrodynamic performance and cause NOISE.

So actually, our sound profile for subs are not realistic.
whether you like it or not ...

nothing better than some links =>

http://www.npmoc.navy.mil/KBay/backgroundnoise.htm

HYDRODYNAMIC NOISE

Self noise resulting from the flow of water past the hydrophones of the hull-borne sonar, its supports, and the hull structure of the platform is hydrodynamic noise. In submarines, this type of noise also includes the noise generated by turbulent pressures upon the hydrophones from flow eddies, as well as vibrations from the submarine's plating and sonar gear. In addition, the water flow around the sonar dome sometimes creates the major portion of self noise. This noise is known as flow induced resonance, or FIR.

http://en.wikipedia.org/wiki/Turbulence

Consider the flow of water over a simple smooth object, such as a sphere. At very low speeds the flow is laminar; i.e., the flow is smooth (though it may involve vortices on a large scale). As the speed increases, at some point the transition is made to turbulent ("chaotic") flow. In turbulent flow, unsteady vortices appear on many scales and interact with each other. Drag due to boundary layer skin friction increases. The structure and location of boundary layer separation often changes, sometimes resulting in a reduction of overall drag. Because laminar-turbulent transition is governed by Reynolds number, the same transition occurs if the size of the object is gradually increased, or the viscosity of the fluid is decreased, or if the density of the fluid is increased.

http://www.cora.nwra.com/~werne/eos/...urbulence.html

and the last (i won't spend all my time to prove you my assertions ...)

http://kr.cs.ait.ac.th/~radok/physics/f2.htm

Turbulent flow

At a critical velocity, a laminar flow becomes turbulent. If for a tube with a circular cross-section the the expression vd/n (v average velocity, d tube diameter) serves as critical number R, the validity of Poiseuille's Law ends at R = 2000, when turbulence begins. In a tube with a diameter of 1 cm, through passes water at 10ºC, R = 2000 yields a mean velocity of about 26 cm/sec. Through this tube, a slower flow, as every flow with a smaller value of R will be laminar (layered). If the velocity increases to this value of if you obtain it by enlargement of the tube diameter or heating of the water (decrease of m), there will occur the change over. The amount of fluid for a given pressure drop will then be smaller than the value, given by Poiseuille's formula, that is, the resistance experienced by the flow increases. The causes of turbulence were not yet known in 1935.


Now we need to determine the critical point for all subs
it's not so difficult to evaluate, as I said before.
And, more difficult, we need to evaluate what happen at the critical point, how much is the gap.
We could talk about that if you have finished to contest physical law :hmm:

Very educational but still not sure how relevant to sim actualite. :roll:

I think OKO has a point about the 'gap' - the changing turbulence characteristics around the hull with speed.
MSGallileos photos of the Virgimia 's wide apperture flank arrays prompted my remark:This may or may not be relevant as the apperture spread may just overcome the barriers of migratory turbulence.

But the manouvering for ''clearing the baffles'' takes on a new meaning when the hull side blind spot moves position.
It means that not only are these 'spots 'transitory but it may mean that certain areas may become relatively
more receptive depending on speed and angle of inclination.

But when one returns from 'Space' the acid test is what can be modelled in the sim and what is ? :hmm:

Correct me if Im wrong, but this argumentation(no air underwater) rules out cavitation, yet it happens. There is always air in water. I guess some "O's" are detaching from the H²O through mechanical force (creating kinda vacuum which then "sucks" the air out of the water, contrary to boiling where the heat is the reason).
So what OKO sais makes alot of sense to me.

Sorry if I'm saying a stupidity, but my English is not so good to follow so complex conversation, but I want to say two things:

Of course there is air in sea water. If not, the fishes couldn't breathe. :yep:

Another thing: the phisycs on the air could not be applied to water. Both are fluids, of course, but liquids (water in this case) are much more dense and complex than the air.
In fact, there no exits a unique formula for the speed of sound in the water. There are one for each place, even for each season, and all this formulas are made by empiric experients, no theorical.

Thats true : if both are fluid, the water have much more viscosity than the air, I can't remember the value exactly but this should be around 100 times.
As you said, most of sound propagation in water are empiric, but there at least a quite known formula to help to compute when turbulence start : it's the Law of Poiseuille, which, coupled with Reynolds number can tell you at what time a hull will start to create turbulence.
The Reynolds number is used to transpose the turbulence effect to a same hull at different scales.

Sound propagation in the water and the time laminar flow become turbulent are 2 different things.
First is mainly from experience (even some formula exist) the other is very well known and ""easily"" calculable (if you have the right data ... including hull size and form and viscosity of the environment)

No, the reason for cavitation is not a chemical reaction where water decomposes (but chemical reactions might happen as a consequence of the high temperature reached when bubbles collapse during cavitation). It simply is that, at high flow speeds, pressure is reduced, so dissolved air will easily go out of solution, and water will boil at lower than normal temperatures. This is also why cavitation is less likely to happen when deep: the pressure is much higher, so (at the same temperature) you need a much faster flow to lower te pressure enough to start the boiling. It obviously also becomes easier with increasing temperature.

Of course, as soon as the bubbles come into a region of slower flow, they are not stable and will collapse. This is noisy and may damage the screw.

Ok...let's not turn this into a flame war. Let cooler heads prevail until Jamie and/or Hutch addresses this issue.