All Posts

A field-ready method to distinguish cold shuts from misruns, trace thermal and filling causes, and test one safe HPDC change without creating porosity.

A field-ready method to distinguish mechanical build-up from metallurgical die soldering, test the right hypothesis, and use foundry AI without trading one defect for another.

Spherical and irregular pores point to different mechanisms. Use morphology, location and process evidence before changing HPDC settings.

Overall accuracy can hide missed rare defects. A practical foundry guide to recall, precision, thresholds, calibration and production validation.

A safe DOE gate helps foundries test parameter interactions before AI recommendations become production settings.

A practical method for proving whether AI-driven cycle-time or energy tuning reduces real energy cost without trading away casting quality.

Why foundries should keep improving after quality is stable—and how one controlled AI-guided parameter change can test cycle-time or energy gains without sacrificing process capability.

A practical validation gate that keeps related shots, dies, cavities, production campaigns and future data out of the wrong train/test split.

A practical measurement-system gate for checking repeatability, reproducibility and label provenance before foundry AI learns from visual, leak, X-ray or dimensional inspection.

A practical method for linking every HPDC shot, cavity and process record to the correct leak, X-ray, dimensional or metallurgical result before AI learns.

A practical governance workflow for detecting model drift, triggering evidence-based retraining and making foundry AI abstain outside validated conditions.

A practical engineering method for building multivariable HPDC process windows so AI recommendations remain bounded, traceable and safe.

A practical six-record data model for building traceable, engineer-validated AI decision support in HPDC and LPDC foundries.

A transparent, modelled HPDC scenario showing how Castella could support porosity diagnosis, parameter review and scrap-reduction decisions.

A practical engineering workflow for setting LPDC lift, filling, holding and release stages using metal level, flow stability, thermal data and defect evidence.

A practical reference for foundry engineers: six common defects in HPDC, LPDC and sand casting, their most likely causes, and the process parameters worth checking first.

Round pore or irregular cavity? Use morphology, CT, shot-profile, vacuum and thermal evidence to separate gas from shrinkage porosity before changing HPDC parameters.

In industrial metallurgy, creating a single perfect aluminum component is relatively simple; producing ten thousand identical components sequentially is an extraordinary engineering challenge. Across both Low-Pressure Die Casting (LPDC) and High-Pressure Die Casting (HPDC) lines, the silent killer of profitability is variance. When thermal drift occurs, the mechanical properties of your castings fluctuate, leading to unpredictable porosity, dimensional warping, and catastroph

For process engineers managing Low-Pressure Die Casting (LPDC) lines, managing the solidification of aluminum alloys like A356 (AlSi7Mg0.3) is a daily thermodynamic battle. With a latent heat of fusion around 389 kJ/kg and a wide solidification range spanning from 615°C to 555°C, maintaining strict directional solidification toward the stalk is incredibly difficult. When this thermal gradient collapses, isolated liquid pools form. The resulting volumetric contraction—typicall

Walk into any active foundry, and you will see digital displays flashing thermocouple readings across the shop floor. Yet, in many operations, these numbers are treated merely as generic safety limits rather than actionable metallurgical data. When a shrinkage defect or a cold shut appears, engineers often scramble to adjust mechanical settings, ignoring the story the temperature curves were trying to tell them. If you are seriously researching how to reduce scrap rate in fou