Account for cosine losses

[dgfl-gh]

Persist the dot product between the engine thrust vector and vessel forward vector to account for cosine losses during persistent thrust

[sswelm]

I think we have to assume a thrust vector always aligns with a part transform up vector. Then its just a matter of taking the cosine of the part up transform against the vessel.

[dgfl-gh]

Yes, should be a safe assumption anyway; never seen an engine (except maybe some VTOL engines that no one is going to use in PT) that didn't have it thrust on its up direction

[dgfl-gh]

Should be working, but still trying to properly test if it is.