Evidence record · UDINUS · SDGs
- Unit
- Rektorat UDINUS
- Period
- 2025
- Goals
- Record
- 0ad3e250-eb1d-4cd7-ac27-8b68ded058fc
- Published
- 2026-10-30
Universitas Dian Nuswantoro conducted research on the detection of scorch occurrence during the production of flexible polyurethane foam. The activity addressed scorch as a problem that consistently jeopardizes foam integrity and resilience. The source states that scorch leads to foam suppression and compactness integrity failure. It also states that the increased likelihood of scorching makes the use of methods intended to avert it important.
The research examined machine-learning-based identification of scorch occurrence during profiling in the flexible polyurethane foam production process. The study utilized a random forest ensemble with feature selection and a data balancing technique to identify production predictors. The source title identifies feature selection and SMOTE-Tomek in relation to a tree-based random forest, while the source text describes the approach as feature selection and data balancing.
The resulting model achieved an accuracy of 0.9998 and an F1-score of 0.9819. The model yielded two distinct cases for non-occurrence and occurrence of scorch, respectively. The ensemble demonstrated that it can effectively and efficiently predict the occurrence of scorch in the manufacturing process for flexible polyurethane foam.
The activity presents research using machine learning, feature selection, data balancing, and a random forest ensemble to identify production predictors and detect scorch occurrence. Its stated focus is the flexible polyurethane foam manufacturing process and the production-related challenge posed by scorch. Further information about the implementation of the data balancing technique, the specific production predictors, the study setting, and the roles of individual researchers is not provided in the source.