FleshJeb

What color is your game clock when this is happening? If it's yellow or red, that means KSP is struggling to do all the physics calculations in real-time.
The KSP engine always does 50 physics updates per game-world second, to ensure the accuracy of the simulation, but that will often take more than one real-world second.
See here: 
You probably want to try sliding the Max Physics Delta-Time per Frame to 0.12. You'll have to get to the settings in the main menu. It's under General, System.
Side note: Until I read the above link, I always thought this affected the quality of the simulation. It does not.

Is it these?
https://bugs.kerbalspaceprogram.com/issues/5250
"When the ore tanks are full and your craft is drilling and converting (via ISRU) at the same time, the drilling proceeds at the maximum rate, even though the utilization of ore may be slower than the extraction of ore. The excess ore is not stored and is simply wasted. This can become a big problem when mining asteroids due to their finite ore. This issue is independent of time warp."
https://bugs.kerbalspaceprogram.com/issues/10125
Partially related to Bug #5250 - the fix left some related issues unaddressed
When ore containers are filled, the actual ore depletion of the asteroid now matches the amount of ore actually being stored (and then processed), however the asteroid drill continues to report 100% (or whatever limited by other factors) load, which also gives full heat production and power consumption way beyound the actual work done by the drill. Even if the drill is actuallly producing nothing, when the storage is full and the converters are off.
In comparison, surface drill reports actual load (and has its power consumption and heat production scaled to it) limited by the ability to store the ore in such situations with full ore tanks, that is being limited by ore conversion rate if ISRU converters are active and completely stopping if no ore is consumed.
I'm big into conservation of mass/momentum in my physics sims. This is one of the many reasons I answer the KSP2 surveys as "Will not buy." I think I mentioned it to you privately, but I'm so done with dealing with poor performance and bugs that I haven't opened KSP since April.

\GameData\Squad\Parts\Misc\PotatoRoid\part.cfg
crashTolerance = 160  

It'll probably only take you 4 hours to build it again, and now you have good motivation to make backups of your craft files. As far as life lessons go, that's a cheap one.
FYI, your craft files are stored in:
C:\Program Files (x86)\Steam\SteamApps\common\Kerbal Space Program\saves\<savename>\Ships\(SPH or VAB)

The aero parts function with the same parameters in water as they do in air. Max lift at 25 degrees AoA, max lift-to-drag at 5-7 for supersonic, 3 degrees for subsonic.
It sounds like your props are sticking partially out of the water to start, and then getting destroyed on impact? KSP has some funny stuff going on with water impact. For a perfect zero-AoA entry they should be able to take 3x their impact rating (15m/s for control surfaces?). I believe this impact resistance gets reduced with the sin of the angle. (I could be wrong, so folks can correct me.)
What I would do is build it with the surfaces flat to the disk, and have them set on deploy at 0 degrees. Start with a fairly low RPM, and slowly increase the deploy angle and RPM until you're moving. (What's the "unstick" speed of KSP water? 0.4m/s?)
If you didn't know, optimal deploy angle depends on the speed the craft is moving. The prop blades are drawing a corkscrew through the air/water, and that corkscrew gets stretched out the faster the craft goes. You want the blades to have an AoA relative to that corkscrew (say 3deg).  Generally speaking, aim for the maximum RPM at all times, and adjust your deploy angle to preserve that.
I haven't done an electric sub yet (I just know entirely too much about KSP physics), so I'd love to see the eventual results.

There's an old bug (that I'm not sure still exists) where making the cargo bay the root part of the craft can prevent it from shielding the parts inside.
Are you closing it with the slider or the button? Even if the slider is set to 0% and the button says open, it's open.
I've never heard of that engine bug, I'm inclined to think it's mod-related. I'm assuming it's a mod engine? Did it work correctly on that planet before?
For diagnosing problems, it always helps to have pictures of the craft and a mod list.
 

@Geschosskopf There's a button in the right-click menu that will let you pick the control axis.

Static/unmoving isn’t as important as low altitude/dense air. They just really thermally inefficient and draggy in the thin hot stuff. Also, I think they only affect core temp directly.
Wings: Since you’are entering belly-down, I was thinking on the dorsal surface with the rear of the wing pointing up. Any orientation that minimizes their drag should do. As long as they stay cooler than the tank.
FYI, in the .cfg files “emissiveConstant” is the thing you’re looking for. 0.95 is really good. I think the default if unlisted is 0.6, but i’d have to look in physics.cfg. Your slant adapters are 0.8. Example: https://wiki.kerbalspaceprogram.com/wiki/Parts/Aero/circularIntake/intakeShockCone.cfg
Likewise, there are parts that make really good insulators when placed in-line. Service Bays are my go-to. My default 1.25m spaceplane stack is nose cone - svc bay - cockpit - precooler.
As usual, I agree with everything @bewing said. Thermal occlusion is mostly voodoo.
Since you’re already invested in RCS Build Aid and EdEx, consider CorrectCoL—If only for the more accurate CoL indicator.
Please ping me when you get this thing working. I’m sure it will be educational and awesome.

I did some testing today with a single-engine Whiplash jet: The radiator wasn't helping much with sustained flight, but if I popped up high or slowed down for a bit, it made a difference. Of note, the one large radiator accounted for about 1/3 of the drag of the entire craft. Also, I think the Precooler may have a much higher overall emissivity than the Mk1 LF tanks. It was frequently exceeding or matching them and the radiator in a number of design options.