A GAN-Augmented Machine Learning Framework for Predicting Raman Characteristics in Carbon Nanofiber Synthesis

This thesis explores a hybrid data-driven framework for predicting the structural quality of Carbon Nanofibers (CNFs) synthesized via Chemical Vapor Deposition (CVD). Building upon prior work employing Conditional Tabular GAN (CTGAN) for data augmentation and XGBoost for quality prediction, this thesis expands the methodology by incorporating: (i) an end-to-end reactor data acquisition system (LabVIEW to cloud-based pipeline), and (ii) a new chemical process modeling approach covering hydrocarbon decomposition, catalyst activation, and CNF growth mechanisms. Contributions include physics-informed feature extraction in syn-thesis phases, synthetic data generation to address small data sets, and the integration of machine learning models to predict the intensity ratio of the D and G peaks in a Raman spectrum (𝐼𝐷 /𝐼𝐺 ), a key indicator of CNF quality. The augmented model achieved 𝑅2 ≃ 0.91, and root mean squared error of 0.065 for 𝐼𝐷 /𝐼𝐺 , outperforming unaugmented baselines (XGBoost 𝑅2 ≃ 0.76; support vector regression 𝑅2 ≃ 0.71). CTGAN preserved data characteristics (≃ 96–97 % marginal fidelity / relationship), allowing robust generalization despite the small-n regime.

Name thesis:degree_name

Master of Science (MSc)

Discipline thesis:degree_discipline

Engineering – Electrical & Computer

Grantor dc:publisher.institution

Graduate Studies

Year dc:date.issued

2026

Author dc:creator

• Foroughi, Amirhossein

Advisor dc:contributor.advisor

• Fapojuwo, Abraham

Committee members dc:contributor.committeemember

• Santos, Ronnie de Souza
• Song, Hua

• Carbon Nanofiber
• XGBoost
• CTGAN

Statement dc:rights

• University of Calgary graduate students retain copyright ownership and moral rights for their thesis. You may use this material in any way that is permitted by the Copyright Act or through licensing that has been assigned to the document. For uses that are not allowable under copyright legislation or licensing, you are required to seek permission.

Language dc:language.iso

en

DOI

https://dx.doi.org/10.11575/PRISM/51023

OAI identifier oai:identifier

oai:ucalgary.scholaris.ca:1880/123989