Comparing case studies of a new structural identification framework based on model falsification reasoning

Abstract eng:
Measurement of infrastructure to improve knowledge of real behavior has the potential to avoid replacement and lower the costs of interventions during service. Building on research into a novel data-interpretation methodology called error-domain model falsification, this paper describes an iterative sequence-free framework for implementation. Two case studies are described to illustrate its use and to demonstrate that the framework is useful in a range of situations for new and old infrastructure elements. In situations where data contains outliers and when modelling is poor, the identification fails through complete falsification of the model class, thereby triggering further iterations to reduce outliers and obtain more knowledge for example, through site visits and sensitivity studies. When there are candidate models, high uncertainty may result in wide prediction bands for performance. Therefore, the usefulness of the results appropriately reflects the quality of the models and of the measurements. This framework has much potential to become a strategic element in the toolbox of the asset manager of the future.

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket