Die casting: die temperature and shot profile set by trial

The thermal balance of the die is unknown

What happens. A die heats up with every shot and is cooled by channels and spray. Thermocouples show a few points. Nobody knows the temperature across the whole cavity, so after a stop, a change of part or a new spray setting the first shots are a gamble.

What it costs. Porosity and cold-shut rejects where the die is too cold or too hot, heat cracks and early die wear, and warm-up shots that are scrapped.

Why the usual fix fails. More thermocouples add cost to an already stressed die and still show points. Setting the cooling by experience works for the part the setter knows. A new die starts the learning again.

How a fixed-price remote diagnosis checks it. From your cycle data (cooling channel flows and temperatures, spray settings, cycle times, die thermocouples) and reject records I look at how the die temperature develops and where it leaves a good window. The fixed-price remote diagnosis tells you whether a computed die temperature can steer the cooling, and what it needs.

Background: Computed measurements

The shot profile stays fixed while the metal varies

What happens. The plunger follows the same speed profile shot after shot. The metal does not: its temperature in the holding furnace, the amount left in the sleeve and the die temperature drift. The cavity fills differently, and the setter reacts to what the rejects show.

What it costs. Rejects from air inclusions and incomplete fills, setter time spent on correcting, and a process that is only as good as the last adjustment.

Why the usual fix fails. A profile made safe for the worst case is slow and cools the metal. Correcting by hand reacts after the bad shots. Fixed limits on the plunger position do not know what the metal is doing.

How a fixed-price remote diagnosis checks it. From shot curves (plunger position, speed, pressure), temperatures and reject data I look at how much the curves vary and which variation coincides with rejects. The fixed-price remote diagnosis tells you whether an adaptive shot control, shot by shot, would pay, and what it requires from the machine.

Background: Learning-augmented adaptive predictive control

Low-pressure casting runs on a pressure curve found by trial

What happens. In low-pressure casting the furnace pressure pushes the metal up into the die. The pressure curve is built by trial for each part, step by step, until the castings are good. Nobody can say if it is the best curve or only one that works.

What it costs. Porosity from a filling that is too fast or too slow, long cycles from careful curves, and trial runs for each new part.

Why the usual fix fails. Copying a curve from a similar part carries over its weaknesses. Slower filling is safer and costs cycle time. Changes in the melt level over a furnace charge shift the result without anyone noticing.

How a fixed-price remote diagnosis checks it. From pressure curves, melt levels, temperatures and quality results I look at which curve parts decide quality and how the melt level changes the filling. The fixed-price remote diagnosis tells you whether an optimised filling curve, computed from a model, can replace the trial.

Background: Trajectory optimisation and setpoint generation