Casting Simulation System ADSTEFAN
"Advanced Defect Prediction Tool (ADPT)" is newly introduced (*1), which realize precise defect prediction using machine learning (*2).
The tool makes "Defect prediction database" which contains some patterns or rules made by machine learning process using training data which contains actual defect information and related analysis results.
The use of the database will significantly reduce a complex and troublesome "fitting" and "interpretation" and will increase the prediction accuracy of defect probability for new products, by automated and objective operation.
Schematics of the ADPT
Example of defect prediction (Cold Shut)
(*1) The ADPT is a limited edition in the ADSTEFVAN Ver.2020, which requires no additional option contract.
(*2) A technology that uses computers to find patterns and rules from a large amount of data, and applies the rules to predict defects for new analyses.
Second rotation axis and integrated movement of multiple parts are available. Combination of "parallel shift" and two "rotation" movements will realize a simulation for special processes such as tilt casting.
Condition 1: Both the mold and the ladle rotate around the 1st axis.
Condition 2: Only ladle rotates around the 2nd axis.
Example of analysis with 2 rotation axis
Ver.2020 clearly indicates the unsigned functions by showing lock mark or inactive indication, which will avoid to proceed further wasteful operation which doesn't work under current contract.
Display example of unsigned functions