Metaization concept for structural health monitoring
Background and motivation
Structural health monitoring (SHM) based on decentralized sensor networks is increasingly drawing research attention. For automated structural health monitoring, interacting (wireless or tethered) sensor nodes, equipped with embedded intelligence, are spatially distributed in civil infrastructure systems. Referred to as “smart structures” or “intelligent infrastructure”, the sensor nodes are capable of autonomously collecting sensor data and, using embedded algorithms, of processing and analyzing sensor data on board in real time. Recent research projects have demonstrated that not only algorithms can be embedded into the sensor nodes; using appropriate decomposition strategies, also models can be embedded into the sensor nodes, such as hybrid, multi-coupled numerical models, which digitally represent the monitored structure in a fully decentralized manner. Thus, it can be expected that next-generation assessment of civil infrastructure is no longer either monitoring-based or model-based. Instead, a new multi-paradigm approach evolves, merging monitoring-based and model-based structural assessment.
Goals and expected results
To precisely evaluate the quality of structural assessment, which is not sufficiently possible using current state-of-the-art methods, this research project aims at developing a metaization concept, resulting in a holistic metamodel architecture. The technological nucleus of the metamodel architecture – the metamodel – is a metalinguistic instrument to be used for holistic modeling and digital representation of structural health monitoring systems (Figure 1). As a result, it is expected that the well-defined formalism provided by the metamodel architecture enables a lifecycle-oriented documentation and a continuous updating of all monitoring-related information. In consequence, the monitoring quality, and thus the quality of structural assessment, can substantially be enhanced. The central question related to the principle of parsimony within SHM, commonly known as “Ockham’s razors”, is to be is to be answered: How complex and resource-consuming must the models be and how simple and resource-efficient can the models be in order to meet the requirements imposed on the quality of structural assessment? To validate the proposed metaization concept, different model classes relevant to typical SHM problems will be considered in a multi-stage validation strategy, using both laboratory tests as well as sensor data taken from civil infrastructure systems in operation.
Professor Dr. Kay Smarsly
Hamburg University of Technology
Institute of Digital and Autonomous Construction