DELORES - A System for Detection and Localization of Structural Damages

Today, structural inspections of large structures, like wind turbines, bridges, etc., are often performed manually by highly trained personnel. Obviously, this inspection approach is both extremely costly and tedious, for which reason this paper provides a presentation of an alternative approach that monitors the structures remotely without human involvement. DELORES (DEtection and LOcalization of RESponse anomalies) is a first version of a system, which utilizes vibration recordings from the healthy and current states of a structure to declare its current health. DELORES employs extended versions of the dynamic damage locating vector (DDLV) and the stochastic dynamic damage locating vector (SDDLV) methods, which include an outlier analysis scheme for a more unambiguous localization of potential damages. The (S)DDLV methods are model-based, which means that a mechanical model of the healthy structure in question is utilized along with measured vibrations of the real structure to determine its current health situation. DELORES provides a user interface with various options that, among other things, make the program suitable for structures modeled in two and three dimensions. The present paper will provide a detailed description of DELORES, along with an application example in which vibrations from a numerical model in three dimensions are treated.