Python in the workflow
ODB/history extraction, data cleaning and feature construction, repeated OOF validation, candidate screening, and FE rerun report assembly.
Engineering & Automation
Engineering code that connects models, data, and decisions.
Programming is presented as an engineering instrument: every example explains its input, operation, output, integration point, and verification method.
Simulation Automation Workflow
Abaqus, control code, and engineering data in one traceable chain.
01Process parameters
02Control histories
03Abaqus/Explicit
04ODB extraction
05Traceable dataset
06OOF + Pareto
07FE rerun
Selected Engineering Work
Three languages, one verifiable workflow.
MATLAB in the workflow
Transforms V1–V8 ring-growth paths into consistent feed-control logic, generates control tables, and performs engineering checks.
Fortran in the workflow
Abaqus/Explicit user-subroutine work for time-dependent process control, presented here through its verified role in the simulation workflow.
01MATLAB control history
02Fortran / Abaqus process control
03Python extraction & model validation
04Independent FE rerun
ML Decision Replay
100 FE cases → repeated OOF → Pareto → independent FE rerun
The interactive, source-data-driven decision replay is kept with the full Research technical evidence.
Collaborative Project
Collaborative Project | ML-Assisted Quality Adjustment Visualization for 20 m-Class Ring Rolling
01Initial condition
Complete first rolling sequence
02ML-assisted prediction and optimization
Parameter-adjustment recommendation
03Corrected condition
Complete second Unity replay
This demonstration is based on 100 thermo-mechanical FE datasets for 20 m-class ring rolling. The two videos show independent, complete processes under different conditions. The transition presents parameter-adjustment recommendations from ML-assisted prediction and optimization; it does not represent reverse deformation of the same ring.
Collaborative visualization evidence; not a substitute for independent FE rerun verification.
Unity collaboration · two independent conditionsMy contribution
Provided and organized the FE data foundation; contributed domain input on ring-rolling motion rules, process constraints, and result interpretation; and supported data mapping, basic Unity operations, and selected C# tasks.
Collaborative development
The Unity visualization architecture and the machine-learning prediction and optimization modules were led by the collaborating developer.