Preface

Process systems engineering is undergoing a fundamental transition. Historically, the field relied almost exclusively on mechanistic models derived from the conservation laws of mass, energy, and momentum. While these deterministic frameworks remain the bedrock of process design and safety analysis, the modern industrial landscape provides an unprecedented volume of high-fidelity operational data. The proliferation of industrial internet of things (IIoT) sensors and advanced distributed control systems (DCS) has created a data-rich environment that traditional modeling paradigms are often ill-equipped to fully exploit.

This book examines the convergence of heuristics, physics, and algorithms. It is not an argument for replacing domain expertise with generic machine learning; rather, it provides a technical framework for integrating the two. Modern chemical engineers must become adept at navigating a spectrum of modeling approaches, from purely mechanistic white-box models to empirical black-box correlations and, increasingly, physics-informed hybrid architectures.

The challenges of industrial data are distinct from those encountered in commercial data science. Process data is characterized by high autocorrelation, multi-rate sampling, and strict physical constraints. A model that predicts a physically impossible concentration or violates a mass balance is of little use in a refinery or a pharmaceutical plant. Therefore, this text emphasizes techniques that respect the underlying physics—such as data reconciliation, domain-aware feature engineering, and Lyapunov-stable control.

The monograph is structured to move logically from the foundations of process data engineering to algorithmic implementation and industrial deployment. It is designed for both academic researchers seeking to apply modern digital tools to chemical systems and industrial practitioners tasked with deploying reliable soft sensors, monitoring systems, and advanced controllers.

The objective is to provide the reader with the theoretical depth to understand these methods and the practical insight to deploy them in complex, real-world environments.