ABSTRACT
Fault diagnosis and maintenance is an important topic both in practice and in research. There is intense pressure on industrial systems to continue reducing unscheduled downtime, performance degradation, and safety hazards, which requires detecting and recovering from potential faults as early as possible. With the exponential growth of monitoring data, fault diagnosis and maintenance face enormous challenges dealing with industrial big data. It is like an iceberg where only a small part of fault information floats on the surface. It is hard to use the previous diagnostic methods to explore the true hidden value. At this time, the problem of transforming the growing volumes of data into the value is a considerable issue. The problem mainly includes two aspects. The first one is how to diagnose and predict failures rapidly or even in real time using novel processing systems. The second one is how to deeply dig out the “big” value of big data by improving the existing methods or leveraging new ones. This section contains three main contents: a method of dimensionless fault eigenvalue extraction, methods for fault diagnosis and fault prediction, and intelligent maintenance under Industry 4.0.
