Could be various reasons.
a.) The air is moving so fast the friction and compression heats it up, resulting in hotter air passing over the radiator. This is why the Concorde stretched 6-10 inches at supersonic flight, due to the intense heating;
D.) the only viable explanation I can provide, is that the fans at higher RPM pump more air into the case than extracting fans can expel, thus trapping warm air inside and subsequently increasing component temperatures. How many intake fans vs. extraction fans do you have?
What temps did you have with original fans? I am not familiar with the Noctua fans, are they designed for use on radiators? or as case fans
So finally my PC build is optimized:
The 8086k was delidded, and before running at 5Ghz temperatures were around 77C when running OCCT - after delidding and running at 5.2Ghz now it runs at 66C. The benchmark test ran for 45minutes, I believe that after all that time the AIO (EVGA 280mm + 2 Noctua Chroma 140mm fans).
I replaced all Noctua PPC3000 fans for the Noctua Chroma 140mm - the noise levels decreased from 46db to 39db.
What I'm concerned and I might ask on other threads is about the voltage on the Vcore - mine is at 1.34V and my system is running at 5.2Ghz and it looks like its stable - so far no blue screens or freezes (after hours and hours of gaming and benchmarks).
Everywhere I look says it needs 1.4V - I have seem other overclocker youtubers using 1.396V as well. Anyway - I think that I'm not sure I understand the relation to the vCore voltage and the multiplier when configuring the BIOS. My multiplier is 52 and my Vcore set to manual at 1.34V however during the OCCT test I can see the voltage gets pushed to 1.36V on the Vcore.
My setup as described is a Asus Maximus Code X and my CPU is a 8086k delidded.