Data-driven innovations in material science and chemical engineering: enabling energy material design and industrial process optimization

Abstract

This thesis explores the integration of data-driven methodologies into material science and chemical engineering, focusing on the design and optimization of energy materials and industrial processes. The research is structured into three interconnected areas: catalyst design, energy material synthesis, and industrial process optimization. Chapter 4 investigates the structure-activity relationship of Pt-based alloys for oxygen reduction reaction (ORR) catalysis using high-throughput density functional theory (DFT) and the SISSO algorithm, demonstrating how computational techniques can predict optimal catalyst candidates. Chapter 5 extends the data-driven approach to optimize the synthesis of lithium iron phosphate (LFP) using machine learning (ML) models, where active learning-based optimization enhances the electrochemical performance of battery materials. Finally, Chapter 6 shifts focus to macro-scale industrial process monitoring, applying long short-term memory (LSTM) networks and multivariate statistical process control (MPCA) for real-time monitoring and prediction of steam boiler operations in industrial settings.
While these three chapters address distinct aspects of material science and chemical engineering, they share a unified methodological framework that employs data-driven techniques to solve complex problems across different scales. From micro-scale catalyst design to material synthesis at the meso-scale and real-time process optimization at the macro-scale, the common philosophy of iterative optimization, integration of computational predictions with experimental validation, and data-driven innovation provides a cohesive strategy. By seamlessly bridging the scales and methodologies, this work demonstrates the broad-reaching impact of data-driven tools in the fourth paradigm of material science and chemical engineering. This thesis highlights the transformative potential of data-driven approaches, underscoring their applicability in accelerating the design of advanced materials, improving process efficiency, and contributing to the advancement of green chemical technology and sustainability.

Date of Award	15 Mar 2025
Original language	English
Awarding Institution	University of Nottingham
Supervisor	Cheng Heng Pang (Supervisor), Kam Loon Fow (Supervisor), Edward Lester (Supervisor) & Siew Shee Lim (Supervisor)