In recent years, driven by the rapid development of electronic, information and communication technologies, the great changes of industrial environment have been witnessed. Due tothe ever increasingdema nds o n product q uality a nd econo mic be nefit, no t onl y the intellige nt co mpo ne nts and devicesare implemented and networked, but also the real-time supervision and controlsystems are running parallel, and the automation degree in modernind ustrial syste ms,such as chemical process, aerospace, nuclear power plant, ship powersystem, transportation s ystem, is continuously growing. Modeling these processes accurately via first principles is becoming time-consuming and impractical, and the supervision and control systems designed based on the process model may only be efficient for a short period in the life-time of the industrial processes.

This challenges and pushes scientists and engineers to develop advanced process monitoring and control methodologies using offline stored or online process data to solve optimal process monitoring and control issues. Therefore, it would be very meaningful if the reliable and efficient process monitoring and control systems could be directly designed based on the i nfor ma tion extracted from the offline stored or online process data. Driven by this reason, the development of advanced data-driven process monitoring and control methodologies is significant in bothresearch and industrial application domains.