Data-driven quality prediction and process optimization for large complex castings

Yutong Guo

doi:10.1016/j.rineng.2026.113269

Data-driven quality prediction and process optimization for large complex castings

2026-10-01

0

OA

AI

Y

Yutong Guo

杨超 (Chao Yang) *

DOI：10.1016/j.rineng.2026.113269

Original

Original request for help

Save

Abstract

En 中文

• Closed-loop workflow links prediction, parameter recommendation, and validation. • Random forest achieved the lowest five-fold CV RMSE among four candidate regressors. • Sectioned PT compares indications before and after process adjustment. • Six adjustable variables were converted into bounded shop-floor parameter ranges. • Field implementation yielded 197 qualified castings out of 200 across four batches.

Keywords:

Casting process

Quality prediction

Production-record label

Scrap-cause screening

Random forest

Deep feedforward neural network

Evolutionary optimization

Process optimization

Data-driven quality prediction and process optimization for large complex castings

Abstract

Journal

Organization

Cited Papers