Advances in Data-Driven Modeling, Fault Detection, and Fault Identification: Applications to Chemical Processes is an accumulation of research on data-driven modeling techniques, and their application towards robust modeling, fault detection and fault identification. The book covers a wide range of basic to advanced empirical techniques in comprehensive detail, and provides a easyto-read guide for academic or industrial researchers that are interested in applying these techniques towards their respective fields. The book starts with exposing the scope of the book, in addition to a brief rundown of the methods discussed, and their importance to academic research and industrial applications. It will also describe some of the chemical processes that will be used to validate and compare the various data-driven techniques, which include the Tennessee Eastman Process and a Fischer-Tropsch bench scale setup. It discusses a first category of the methods, that covers basic and advanced robust empirical techniques, followed by a second category of the methods discussed, that covers prominent empirical statistical charts used to detect faults in multivariate systems, and finally a third category of the methods, that covers conventional and novel multiclass classification machine learning techniques that can be used to accurately differentiate in batch or real-time between different fault classes in industrial process or academic applications.

• Bridging the gap between experts and beginners by delving into the necessary mathematical formulation for experts and simplified linguistic interpretation and analysis for beginners
• Provides detailed guidance on optimization and/or tuning of empirical methods towards a specific objective, facilitating reproduction of results in order to verify their accuracy
• Provides guidelines on the appropriate application of each method, allowing to maximize the potential benefit of data-driven techniques quickly without wasting time and effort exploring different techniques