Abstract and subjects
Structural health monitoring (SHM) is a technique that assesses engineering structures online in order to identify damage and ultimately predict their remaining useful life. SHM therefore enables a more efficient maintenance and operational decision-making process. Traditionally, SHM has been focussed on a single structure or system. In most maintenance strategies, detected damages or defects are repaired before they can progress further. However, it is rare that one would have access to measurements from all possible damage scenarios, and particularly rare that one may have measurements of multiple site damage. As a result, studying single structures can give a limited view of the variety of damage that is possible. Population-based SHM aims to extend the effectiveness and applicability of SHM to “populations” of structures by sharing information between them, allowing engineers to obtain a much broader understanding of the damage that can occur in all structures. This paper discusses the most important aspects of using databases in population-based SHM and will also focus on the exploitation of the unique Echo framework, providing a platform for diagnostics across populations of wind turbines. The focus is on forming a data store of features using a coherent, semantically rich labelling procedure which describes the characteristics of both the features and the structures from which they were acquired. By doing so, the aim is to improve the ease of access to informative and useful data. Gaussian process regression to predict the environmental and operational variation across a population of structures is employed here as an illustrative example of the uses of a database framework.