Dynamic Complex Event Processing for Industrial Monitoring Systems

Abstract

Using Complex Event Processing (CEP) as part of monitoring systems is a state-of-the-art approach in the manufacturing industry that still requires development. The industry is increasingly moving towards implementing Service Oriented Architecture (SOA) based systems to respond to increasing demands of interoperability amongst other operations in a business organisation. Complex event processors are used as part of monitoring systems but current complex event processors are usually system specific. This thesis aims to propose and demonstrate a more dynamic approach for implementing an industrial monitoring system using complex event processing. Service Oriented Architecture uses event-based messaging to communicate between different devices and systems. This creates large amounts of data in the monitored system. In order to infer important information from this vast body of data the CEP is used to query through the events. These queries are predefined and cannot be changed during runtime. The CEP holds the main logic of the monitoring system and thus dictates what the system actually monitors. Monitoring system requires the possibility to change the monitoring logic. This is why a method of dynamically adding queries will be proposed in this thesis. In order for a SOA-based monitoring system to be dynamic the CEP needs to be dynamic. This thesis proposes a CEP solution with generic implementation, dynamic query definition during runtime and the possibility to use recursive user defined functions that allow reusing query templates in different solutions. The developed CEP is tested with two different implementation use cases. First one a simulated use case that tests the monitoring system performance with large amounts of events. Second one a manufacturing line implementation to demonstrate the monitoring system in an actual manufacturing environment. Tests were run on both use cases to gain information on how the CEP performs and to demonstrate the functionality of the developed monitoring system. The developed CEP was used as a part of oil lubrication use case for IMC-AESOP project. IMC-AESOP project was an EU project researching how to apply state-of-the-art SOA-based systems to the industrial automation field.