Prevent LEG from drawing significantly higher amounts of coolant than expected

[ReaganAR]

Fixes issue GTNewHorizons/GT-New-Horizons-Modpack#14104 by moving getPerAspectEnergy call into EU generation loop.

As an example, some sample data from supplying Ira with cryotheum as the coolant, over 60s+:

Tier	Expected Coolant Draw	Current Coolant Draw	This PR
HV 4A	0.9503	200.83	0.944
EV 4A	3.8014	203.27	3.8282

And with Adept hatches instead of novice:

Tier	Expected	Current	This PR
EV 4A	1.907	200.42	1.99
IV 4A	7.6028	201.878	7.802

[AbdielKavash]

Line 437-443:

            if (EUt == 0 && aspects.size() != 0) {
                if (!isValidEssentia(aspects.getAspects()[0]) || getPerAspectEnergy(aspects.getAspects()[0]) == 0)
                    continue;
                EUt += getPerAspectEnergy(aspects.getAspects()[0]) * mStableValue / 25;
                aspects.reduce(aspects.getAspects()[0], 1);
                if (aspects.getAmount(aspects.getAspects()[0]) == 0) aspects.remove(aspects.getAspects()[0]);
            }

This section should be completely removed. If the code ever gets here with EUt == 0, this means that the previous loop did not find any valid essentia in the input hatch. Trying the zeroth essentia again won't help, it calls the problematic getPerAspectEnergy again (several times), and is otherwise useless and confusing.

---

Ideally I would prefer a rewrite to use ProcessingLogic or similar (I am not sure what state essentia support for this is in), but after addressing the comment below this works as a quick fix.