Condition Monitoring (CM) is an extremely critical application of the Internet of Things (IoT) within Industry 4.0 and Smart City scenarios, especially following the recent energy crisis. CM aims to monitor the status of a physical appliance over time and in real time in order to react promptly when anomalies are detected, as well as perform predictive maintenance tasks. Current deployments suffer from both interoperability and management issues within their engineering process at all phases– from their design to their deployment, to their management–, often requiring human intervention. Furthermore, the fragmentation of the IoT landscape and the heterogeneity of IoT solutions hinder a seamless onboarding process of legacy devices and systems. In this paper, we tackle these problems by first proposing an architecture for CM based on both abstraction layers and toolchains, i.e., automated pipelines of engineering tools aimed at supporting the engineering process. In particular, we introduce four different toolchains, each of them dedicated to a well-defined task (e.g., energy monitoring). This orthogonal separation of concerns aims to simplify both the understanding of a complex ecosystem and the accomplishment of independent tasks. We then illustrate our implementation of a complete CM system that follows said architecture as a real Structural Health Monitoring (SHM) pilot of the Arrowhead Tools project, by describing in detail every single tool that we developed. We finally show how our pilot achieves the main objectives of the project: the reduction of engineering costs, the integration of legacy systems, and the interoperability with IoT frameworks.
