کد مقاله کد نشریه سال انتشار مقاله انگلیسی نسخه تمام متن
560708 1451889 2012 14 صفحه PDF دانلود رایگان
عنوان انگلیسی مقاله ISI
Identifying damage locations under ambient vibrations utilizing vector autoregressive models and Mahalanobis distances
موضوعات مرتبط
مهندسی و علوم پایه مهندسی کامپیوتر پردازش سیگنال
پیش نمایش صفحه اول مقاله
Identifying damage locations under ambient vibrations utilizing vector autoregressive models and Mahalanobis distances
چکیده انگلیسی

This paper presents a study for identifying damage locations in an idealized steel bridge girder using the ambient vibration measurements. A sensitive damage feature is proposed in the context of statistical pattern recognition to address the damage detection problem. The study utilizes an experimental program that consists of a two-span continuous steel beam subjected to ambient vibrations. The vibration responses of the beam are measured along its length under simulated ambient vibrations and different healthy/damage conditions of the beam. The ambient vibration is simulated using a hydraulic actuator, and damages are induced by cutting portions of the flange at two locations. Multivariate vector autoregressive models were fitted to the vibration response time histories measured at the multiple sensor locations. A sensitive damage feature is proposed for identifying the damage location by applying Mahalanobis distances to the coefficients of the vector autoregressive models. A linear discriminant criterion was used to evaluate the amount of variations in the damage features obtained for different sensor locations with respect to the healthy condition of the beam. The analyses indicate that the highest variations in the damage features were coincident with the sensors closely located to the damages. The presented method showed a promising sensitivity to identify the damage location even when the induced damage was very small.


► We present a sensitive method to identify damage locations in steel beams.
► An experimental study was performed to validate the method under ambient vibrations.
► Vector time series models were fitted to the measured vibration responses.
► The statistical variations in the models are used to identify the damage locations.
► The damage locations were identified even when the damage level was very small.

ناشر
Database: Elsevier - ScienceDirect (ساینس دایرکت)
Journal: Mechanical Systems and Signal Processing - Volume 26, January 2012, Pages 254–267
نویسندگان
, , , ,