Good data from both of you. We track energy per-part on 6 presses and here are a few things that surprised us:
The biggest variable isn't the forming force — it's the slide velocity profile. We had two jobs on the same 300T press, similar tonnage, but one used a fast approach / slow press / fast return profile and the other used constant speed throughout. The optimized profile used 18% less energy per part because the motor wasn't fighting inertia as hard during the working stroke.
On the regen side — we measured actual recovery on our Aida DSF-N1 at about 8-12% of the stroke energy, depending on the return speed. The capacitor bank helps but honestly the bigger win was just reducing unnecessary slide travel. We had 20mm of extra approach distance on most jobs just from lazy programming. Cutting that saved more energy than the regen recovered.
For anyone trying to calculate cost per part: don't forget the idle draw. Our 400T pulls about 3.5 kW just sitting there with drives enabled. If you're running 60% uptime on a shift, that idle power adds up to almost 30% of your total energy bill for that press. We now auto-disable drives after 5 minutes of no cycle signal — saved us about $2,100/year on that one press alone.