Towards Population-Based Structural Health Monitoring, Part II: Heterogeneous Populations and Structures as Graphs

Abstract

Information about the expected variation in the normal condition and various damage states of a structure is crucial in structural health monitoring. In an ideal case, the behaviour associated with each possible type of damage would be known and classification would be possible. However, it is not realistic to obtain data for every possible damage state in an individual structure. Examining a population of structures gives a much larger pool of data to work with. Machine learning can then potentially allow inferences across the population using algorithms from transfer learning.The degree of similarity between structures determines the level of possible knowledge transfer between different structures. It is also useful to quantify in which ways two structures are similar, and where these similarities lie. This information determines whether or not certain the transfer learning approaches are applicable in a given situation. It is therefore necessary to develop a method for analysing the similarities between structures. First, it must be decided which properties of the structure to use when measuring the similarity. For example, comparing 3D CAD models or Finite Element models is not a suitable approach, since these contain a lot of irrelevant information. It is better to abstract this information into a form that contains only the relevant information. This paper proposes Irreducible Element (IE) models, which are designed to capture the features that are crucial in determining whether or not transfer learning is possible. This information is then converted into an Attributed Graph (AG). The Attributed Graph for a structure contains the same information as the Irreducible Element model; however, the graph carries this information as a list of attributes attached to nodes. Organising the information in this manner makes it easier for graph-matching algorithms to perform a comparison between two structures. This comparison can then be used to generate a measure of similarity between the two structures and determine the most appropriate transfer learning method.

Towards Population-Based Structural Health Monitoring, Part IV: Heterogeneous Populations, Transfer and Mapping

“…A population, in the context of PBSHM, is a group of structures (the smallest being a group of two structures) that provide information required for performing health monitoring. This general definition of a population can be further divided into two categories: homogenous and heterogeneous populations [3]; these groupings relate to the level of dissimilarity within a population, where both population types benefit from knowledge transfer. Colloquially, a homogeneous population is one in which each structure in the population can be deemed nominally identical for a given context [2,3].…”

“…The concept of PBSHM is to incorporate the feature and label data from each aeroplane to generate a machine learning-based approach that generalises across the complete fleet for all damage scenarios, especially when many members of the fleet have no labelled data associated with them. This work is Part IV of a series of papers on PBSHM [2,3,4,5,6], where for an overview of PBSHM and further motivation the reader is referred to earlier parts [2,3,4].…”

“…These constraints can be limiting in the context of PBSHM. With reference to the topology of a graphical representation of a structure (discussed in Parts II and III of the paper series [3,4]), this paper presents a mathematical underpinning for when domain adaption is possible within the context of PBSHM.…”

Towards Population-Based Structural Health Monitoring, Part III: Graphs, Networks and Communities

“…The degree of similarity between two structures will determine the information transfer possible. In order to calculate the degree of similarity between structures, it is first necessary to create an Irreducible Element (IE) model of the structure (as described in [7]) and then convert this IE model into an attributed graph (AG) [8]. This paper will describe how the AGs from two separate structures are compared in order to determine the degree of similarity.…”

An Ontological Approach to Structural Health Monitoring

“…Although applicable to SHM in a broad sense, an ontology could be particularly useful in a Population-based SHM (PBSHM) setting [4,5,6,7,8], where the goal is to develop general inference tools across a population. Here an ontology would identify useful, and perhaps overlooked, connections between objects such as Irreducible Element (IE) models and Atribbuted Graph (AG) representations of structures [5,6] and appropriate knowledge transfer methods like transfer learning [7] and 'forms' [7,4]. This may provide benefits in highlighting appropriate methods for each data source such that destructive phenomenon like negative transfer in transfer learning are avoided.…”