Methodology for Condition Monitoring of Industrial Systems Based on Edge Computing

Abstract

Industry in the modern sense brings with it challenges in many areas. Among these challenges is the predictive maintenance of industrial automation equipment. Induction motors, industrial robots and more are now commonly used in manufacturing processes around the world. The concept of predictive maintenance has its roots in the 1990s and there are many research papers and publications in this direction. Manufacturers are also trying to reduce production costs to a minimum as part of competition and price battles, which leads to the requirement for an overall increase in the efficiency of the production process. In the last few decades, the industry has been undergoing total digitalization and there is also a more efficient handling of data within Cloud and Edge platforms. This qualification thesis aims to present a methodology to monitor the condition of industrial machines commonly used in modern industry. It also defines an basic Edge-enabled framework for Condition Monitoring and Predictive Maintenance (PdM). It also discusses modern approaches to data handling for predictive maintenance needs within Edge platforms and describes the broader context of this issue.