Our solution: DropForm

Fossil plastics have had 70 years of factory know-how built up behind them. Bio-based plastics are starting from scratch. DropForm uses AI, trained on real production data, to give them that head start, so factories can run them with far less trial and error.

Know-how is data nobody wrote down

Every job a factory runs leaves a record behind: which material, which mould, which machine, what the settings were set to, and whether the parts came out good.

For fossil plastics, seventy years of those records have quietly turned into instinct. For bio-based plastics they exist too. They are just scattered and thin, and nobody has put them together.

0production runs, and counting

Settings that work are not random

Plot enough past runs and the successful ones sit together. For any given material, product and machine there is a narrow band of settings that works. The catch is that the band moves every time one of those three changes.

DropForm learns where the band sits, including for combinations nobody has run before.

Each dot is one past production run. The dark ones produced good parts. The shaded box is where DropForm predicts the good settings will be.

The factory gets a starting point

Not a report to interpret. Four numbers to set on the machine before the first attempt, for the exact material and product in front of them.

Melt temperature0

Mould temperature0

Hold pressure0

Cooling time0

How hot to melt the plastic, how hot to keep the mould, how hard to push it in, how long to let it cool. Example output for a thin food tray on a mid-sized machine.

First attempt, not the twelfth

Trial and error does not disappear. It gets short. One adjustment, and the second attempt is a part you can sell.

Without DropForm12 attempts

With DropForm2 attempts

0attempts

0days of machine time

0of wasted material

Seventy years of head start, without waiting seventy years for it.