[WIP] thermocline refraction effect

[castorp345]

in sim.cfg the value ranges given for thermal layer signal attenuation are, for passive sonar, "3 equals signal reduction to 33%", and for active sonar "5 equals signal reduction to 20%". assuming the sound energy source is proceeding from 45deg, it appears that a 'nominal' median for the angle of refraction should be c22.5deg (see == http://www.eugeneleeslover.com/USNAVY/CHAPTER-28-A.html and the diagram under section 28A4 "The Limitations of Sonar"). since attenuation corresponds to the angle of refraction i'd suggest using the same value--22.5deg--to determine at least the passive sonar's signal attenuation value: 1.7032 (as opposed to the original sim.cfg's '3'=33%).
active sonar is a little more difficult and i'd certainly welcome further input here... i'm assuming that the higher energy level of a focused 'ping' results in less apparent attenuation, but the question is to what order?

[edit: fixed botched math :oops:]


:lurk:

[Bill Nichols]

Of course, active sonar has to cross the thermocline twice, so the amount of attenuation is compounded compared to the passive case.

[castorp345]

Of course, active sonar has to cross the thermocline twice, so the amount of attenuation is compounded compared to the passive case.
of course! :oops:

thanks Bill!

:D

[castorp345]

ok, after some further examination i've come up with what seems to be more "reasonable" numbers:

[edit: probably not :nope: , still tbd]

for passive, Thermal Layer Signal Attenuation=1.7032 ; where '1' means no signal reduction and '3' equals signal reduction to 33%, then '1.7032'=signal reduction by 22.5% (ie the angle of refraction assuming a 45deg initial source)

for active, Thermal Layer Signal Attenuation=2.0078 ; where '1' means no signal reduction and '5' equals signal reduction to 20%, assuming a HF pulse of 35KHz and max 'nominal' emmissions of c20KHz gives a modifier of 1.75 which, when compounded (to account for crossing the thermocline twice), produces '2.0078'=signal reduction by 19.69% (apparent attenuation)

(if anyone has more accurate Hz values for japanese sonar of this time period i'd love to know!)

[PeriscopeDepth]

Japanese sonar frequencies can be found in here http://www.fischer-tropsch.org/primary_documents/gvt_reports/USNAVY/USNTMJ%20Reports/USNTMJ-200B-0343-0412%20Report%20E-10.pdf

PD

[castorp345]

thanks PD!

[PeriscopeDepth]

thanks PD!

No credit to me, was posted in the ASW research thread by someone else.

PD

[castorp345]

:oops:
'seems i need to be more attentive...

thanks for calling it to my attention though!
:D

[Dr.Sid]

I have studied this subject quite a lot and I have written some simulation .. anyway I'm no pro.
20% is pretty reasonable value for common conditions. But it depends on angle and very much aspects.
Also variable speed of sound can greatly increase the signal at some locations (several times louder) by focusing sound 'beams' .. two most common situations are convergence zones and deep sound channel. The difference between the best and worst can be really thousands of percent.

For active sonar the loss should be almost perfect double. As SubGuru said.
The situation from under the layer and from above the channel is very different, but if you take 2 points in the sea the sound transfer from A to B is the same as from B to A. Also active has some other effect which decreases it's performance, especially reverbation, which is reflection from all that small particles in the sea. There is nice example with light reverbation .. while you can see other light sources in fog quite well, it's really hard to light something in fog with strong lamp. All you will see will be just the fog. Sound reverbation is just that. It is that very long slowly weakening echo after the ping.

Here are some images from my simulator. Brightness shows 'loudness' of the sound at that position, this is side view, distance is on horizontal axis, depth is on vertical.
Please not that sound bending effect is highly increased so it can be visible on such a small scale. The blue line shows sound speed profile. These are not typical, but best for the demonstrated phenomenon.

http://roger.questions.cz/other/profiles.html

[castorp345]

thanks Dr. Sid!

i just replied somewhat at length in the thread on the dw forum (http://www.subsim.com/radioroom/showthread.php?p=475007&posted=1#post475007), but i'm very interested in your observation that "20% is a pretty reasonable value for common conditions"... can you venture what sort of spread around that 20% might still be considered to fall within nominal values? what about the case of passive sonar? would that have the same set of probabilities given "common conditions"?
also, doesn't the higher frequency of active sonar emissions in their end result somewhat compensate for the loss from crossing the layer twice? (and all this is of course strictly for our purposes here from the standpoint of a surface dd and its subsurface prey below the layer.)

many many thanks!

hc

[Dr.Sid]

Very low frequencies are affected less by the layer. I mean frequencies where wavelength is same size or more than layer depth. When the wave does not 'fit' above the layer the sound is not affected that much. With layer 100m deep we are talking about frequencies about 15 Hz and less, with deeper layers even lower frequencis.
All frequencies above that are affected more or less same .. they are bend (as pictured on the pictures). Layer at shallow angles can reflect practically all the sound back to the surface. Passive or active, noise or tonals, it does not matter.

With concrete frequencies (as used in active sonar) and calm water there is yet other strong effect which can practically cancel echo of contacts at some locations. It is called Lloyd mirror and it is caused by interference of direct and surface reflected sound waves. As the target get closer to the pinger, it periodically vanish and appears again.

As for the 20% .. I mean at some situations this is even too optimistic. It's by no mean too strong ettenuation.
If SH4 uses simple 'attenuation of sounds below the layer' model, it's so simplistic you wont get realistic results from it anyway.

I'll try to make the sound propagation simulator little more user friendly and I'll release it so you can play with it.

[castorp345]

thank you again for the wonderful information and taking the time to inform, Dr. Sid!
just to be clear though (as i'm feeling still a little confused by this), in sh4's simplistic model, that '20%' is the final value of the probability (ie a reduction by 80%)... so you're suggesting then that attenuation can cause anywhere up to 100% signal loss under 'normal' conditions (as in the case you mention of shallow angles)? this certainly seems to be born out by even just that first image from your simulator you posted (barring that slight scatter immediately below the emitter)... but i want to make sure that i've got my head on straight with this. :oops:

also, i'd love to play around with your sound propagation simulator (the images you shared look terrific), so thanks as well for that! does it give numeric output as well as visual representation? what sort of initial-state parameters are configurable? 'looks great!

cheers
hc

[Bill Nichols]

I was looking for some useful information to help answer your question, and ran across this nugget in "Principles of Naval Weapons Systems" by David R. Frieden:

"The [submarine's] optimum approach for close approach to a target with minimum probability of counter-detection is approximately

Best Depth = 17 x SQRT(Z)

where Z is the layer depth in meters.

This is accurate down to a layer depth of 60 meters. Below that, the best depth for approach is a constant 60 meters below layer depth."


I doubt that this holds true for SH4 (unless they've modeled sonar layers with more fidelity than I believe), but it's interesting none-the-less.

:know:

[Dr.Sid]

I pretty improved the sound propagation simulator during the weekend and I hope I'll release it this week.

[Dr.Sid]

Ok .. so here is the first version. Please be kind if it crashed once or twice :|\\

http://roger.questions.cz/other/SoundPropagation.1.0.zip

[castorp345]

thanks Dr. Sid!
'looking forward to playing with it! :D

in the meantime, i've got my partner-in-crime from my other big 'mod' project (http://www.orbitersim.com/Forum/default.aspx?g=posts&t=6480) working on some code for an external app to help massage the thermal attenuation values in SH4...
more details to follow soon and many thanks to all for the very helpful info!

cheers
hc

[badhat17]

I feel that some time should be spent on researching the depth that thermocline layers are found in the relevant op areas and the frequency of occurrence also. Ingame layers are found quite often and usualy around 150 feet.
Is that realistic I wonder ?
I think it is quite likely that the devs made layers a frequent occurrence in game as they are probably fed up with hearing players constantly asking " what about the layers? " each time a new sim is released.
If a realistic average working layer effect is obtained by tweaking then it may have an impact which is far from historical if the layers are encountered too frequently.

[castorp345]

I feel that some time should be spent on researching the depth that thermocline layers are found in the relevant op areas and the frequency of occurrence also. Ingame layers are found quite often and usualy around 150 feet.
Is that realistic I wonder ?
no, it's not really realistic.
the layer depth does vary considerably over the course of the year (in the equatorial eastern pacific for instance depths have been reported as ranging from 25m to 115m)...
while it doesn't seem as though one can change the layer depth in SH4, one can adjust the attenuation value (ie the effect the layer has on signal detection probability), and here's a little app to help with that ==> http://www.subsim.com/radioroom/showthread.php?p=481302#post481302 :D

this works by adjusting the attenuation value by season (simulating the varying depth of the thermocline over the course of the year), as well as introducing a degree of randomness to the effect (thereby accounting in some measure for the broad spectrum of possible conditions that affect detection probabilities for contacts across the layer).


I think it is quite likely that the devs made layers a frequent occurrence in game as they are probably fed up with hearing players constantly asking " what about the layers? " each time a new sim is released.
likely.

If a realistic average working layer effect is obtained by tweaking then it may have an impact which is far from historical if the layers are encountered too frequently.
try the little app out and see what you think!

[badhat17]

It would likely be a waste of time my testing your app castorp as I am the kind of sub skipper who would bring dishonour to the service, bring me your weak, your old, your defenseless merchants. I have yet to be pinged in my campaign and long may that continue.

[castorp345]

It would likely be a waste of time my testing your app castorp as I am the kind of sub skipper who would bring dishonour to the service, bring me your weak, your old, your defenseless merchants. I have yet to be pinged in my campaign and long may that continue.
hehehe

you should check out post #116 in this thread ==> http://www.subsim.com/radioroom/showthread.php?t=108688&page=6

[badhat17]

Perhaps I should have stated that I have yet to see a destroyer in my campaign, I will take easy pickings whenever they are offered.