ESTIMATION OF THE DEFECT HAZARD CLASS IN BUILDING STRUCTURES: A DECISION SUPPORT SYSTEM

Abstract

Technical condition monitoring of building structures located on hazardous facilities is a necessary requirement for their sustainable functioning. In this regard, the problem of development intellectual monitoring systems that allow to detect and classify operating defects by the hazardous level becomes very urgent. The study presents an approach of building decision support system (DSS) for detecting defects in building structures and estimation of their hazard class. Proposed approach is based on multi-criteria assessment of consecutive measurements acquired by acoustic emission method. A distinctive characteristic of the proposed approach is the ability to take into account the evolution of defects by mapping each AE time-series to diagnostic features matrix and analysing these matrices in sliding windows with overlay. Each matrix is validated by two criteria that form the necessary and sufficient conditions of the existence the evolving defects in building structure. They include the criterion for changing the number of clusters and the criterion for changing the acoustic emission activity. Proposed method was verified on the experimental data acquired from the technical condition monitoring of the vertical oil tanks. The results obtained from the experiment confirm the proposal that this approach can be utilized for effectively solving the problem of conditional monitoring of building structures located on the hazardous facilities allowing to detect and classify defects by their hazardous level.

Keywords

• decision support system, acoustic emission, feature extraction, clustering, defect hazard class estimation, image processing, distance transform