Systematic Mapping Studies and Prototyping on Edge Computing for Condition Monitoring

Edge computing brings computation close to the source of data, offering computing models for terminal devices and data, keeping them secure and private. This is achieved through local processing and storage at the network's edge, which minimises energy consumption, costs, and network traffic. As a result, data intended for transmission to cloud services is reduced in size, requiring minimal computing resources. This is significant in areas with limited data connectivity where edge computing plays a crucial role in performing data analysis and root cause detection to prevent damage or destruction of mechanical equipment. The first part of this thesis focuses on assessing the current status of research in edge computing systems through a systematic mapping study. The literature on edge computing resulted in efforts on task scheduling, power optimisation and data management solutions. To identify edge solutions relevant to the maritime field, an additional systematic mapping study was conducted revealing conceptual models intended for real-life implementation. The research raised solutions for monitoring and communication functions for maritime vessels. The latter of this dissertation contains research on developing an edge analytics solution designed for autonomous devices and components, with an emphasis on reliability and availability. This involves the exploration of raw data processing algorithms, edge processing requirements, and the integration of cloud and operator infrastructure services. Edge analytics adds predictability, enabling operational optimisation, diagnostics, and predictive maintenance. An experimental system was created to understand an effective edge architecture, assess its reliability, and extract insights from data. This edge system was tested in a laboratory environment and later deployed on a cruise ferry that was travelling in the Baltic Sea. Over a period of approximately ten months, the edge system served as a tool for monitoring engine vibration data of the cruise vessel. The third segment of this thesis focuses on an exploration of the feasibility of data exchange and the advantages it brings to various stakeholders from their engagement in digital centralised partnerships. Finnish enterprises were the subjects of semi-structured interviews, having the objective of comprehending their collaborative practices with respect to data sharing, as well as identifying any impediments hindering these processes. Insights were gathered on legal, economic and technical requirements and possibilities for performing data sharing. In summary, this research includes a review of the state-of-the-art edge computing methodologies, a proof of concept implementation of edge vibration analysis, and an evaluation of data-sharing solutions.