Intelligent Predictive Maintenance Frameworks for Fault Classification

With the rising demand for complex, integrated and autonomous systems in the field of engineering, efficient and versatile Predictive Maintenance (PdM) frameworks have become a requirement for monitoring the health status of these systems since safety, reliability and optimum asset utilisation, are key issues. However, due to the continuously changing dynamics of industrial operations, the data recorded for developing PdM frameworks are often high-dimensional and characterised by undesirable features such as high level of uncertainty, class imbalance and multiclass among others. These undesirables limit the efficiency of existing PdM frameworks in producing desirable results. For these reasons, this book has proposed three hybrid and novel PdM frameworks capable of handling such undesirable features through the hybridisation of machine learning techniques. The proposed hybrid frameworks advance the field of PdM by improving the accuracy of fault diagnosis as the issue of undesirable features impedes the ability of machine learning algorithms to produce desired results.

Autorentext

Dr. A. Buabeng is a data scientist with interest in the application of machine learning in predictive maintenance. A. Simons is a Professor of mechanical engineering with interest in the design of machine elements and maintenance engineering. Dr. Y. Y. Ziggah is a researcher with interest in engineering application of artificial intelligence.