Machine Learning in Python for Process Systems Engineering: Achieving operational excellence using process data

focussed on the key steps in soft sensor design: data selection; model structure selection and model validation, respectively. Extensions to the basic steps of soft sensor design, namely soft sensor performance enhancement and the modifications needed to facilitate different industrial process applications follow in Chapter 7 and 8, respectively. Widening the applications range and role of soft sensors to fault detection and sensor validation configurations is dealt with in Chapter 9. A great strength of Soft Sensors for Monitoring and Control of Industrial Processes is the use, throughout the text, of a set of industrial case studies to demonstrate the successes and drawbacks of the different methods used to create soft sensor models. A number of different methods may be used in each separate step of the soft sensor design process and the industrial case studies are often used to provide explicit comparisons of the performance of these methods. The industrial control and process engineer will find these comparison exercises invaluable illustrations of the sort of results that might be found in industrial applications. The monograph also highlights the importance of using knowledge from industrial experts and from the existing industrial process literature. This is an important aspect of industrial control that is not very widely acknowledged or taught in control courses. Most industrial processes have already generated a significant experimental knowledge base and the control engineer should develop ways of tapping into this valuable resource when designing industrial control schemes.