تجزیه و تحلیل حساسیت از شناسایی آسیب ساختمان تکمیل شده با استفاده از نمونه سازی مجازی

The timely and accurate assessment of the damage sustained by a building during catastrophic events, such as earthquakes or blasts, is critical in determining the building's structural safety and suitability for future occupancy. Among many indicators proposed for measuring structural integrity, especially inelastic deformations, Interstory Drift Ratio (IDR) remains the most trustworthy and robust metric at the story level. In order to calculate IDR, researchers have proposed several nondestructive measurement methods. Most of these methods rely on pre-installed target panels with known geometric shapes or with an emitting light source. Such target panels are difficult to install and maintain over the lifetime of a building. Thus, while such methods are nondestructive, they are not entirely non-contact. This paper proposes an Augmented Reality (AR)-assisted non-contact method for estimating IDR that does not require any pre-installed physical infrastructure on a building. The method identifies corner locations in a damaged building by detecting the intersections between horizontal building baselines and vertical building edges. The horizontal baselines are superimposed on the real structure using an AR algorithm, and the building edges are detected via a Line Segment Detection (LSD) approach. The proposed method is evaluated using a Virtual Prototyping (VP) environment that allows testing of the proposed method in a reconfigurable setting. A sensitivity analysis is also conducted to evaluate the effect of instrumentation errors on the method's practical use. The experimental results demonstrate the potential of the new method to facilitate rapid building damage reconnaissance, and highlight the instrument precision requirements necessary for practical field implementation.

Rapid and accurate evaluation approaches are essential for determining a building's structural integrity for future occupancy following a major seismic event. The elapsed time could translate into private financial loss or even a public welfare crisis. Current inspection practices usually conform to the ATC-20 post-earthquake safety evaluation field manual and its addendum, which provide procedures and guidelines for making on-site evaluations [1]. Responders such as ATC-20 trained inspectors, structural engineers and other specialists conduct visual inspections and designate affected buildings as green (apparently safe), yellow (limited entry), or red (unsafe) for immediate occupancy [2]. The assessment procedure can vary from minutes to days depending on the purpose of evaluation [3]. However it has been pointed out by researchers [4] and [5] that this approach is subjective and thus may sometimes suffer from misinterpretation, especially given that building inspectors do not have enough opportunities to conduct building safety assessments and verify their judgments, as earthquakes are infrequent. Despite the de-facto national standard of the ATC-20 convention, researchers have been proposing quantitative measurement for more effective and reliable assessment of structural hazards. Most of these approaches, especially non-contact, build on the premise that significantly local structural damage manifests itself as translational displacement between consecutive floors, which is called interstory drift [6]. Interstory drift ratio, which is interstory drift divided by the height of the story, is a critical structural performance indicator that correlates the exterior deformation with the internal structural damage. The larger the ratio is, the higher the likelihood of damage. For example, a peak interstory drift ratio larger than 0.025 signals the possibility of serious threat to human safety, and values larger than 0.06 translate to severe damage [7]. This research proposes a new approach for estimating IDR using an Augmented Reality (AR) -assisted non-contact method. AR superimposes computer-generated graphics on top of a real scene, and provides contextual information for decision-making purposes. AR has been shown to have several potential applications in the civil infrastructure domain such as inspection, supervision, and strategizing [8]. AR-assisted building damage detection is a specific type of inspection.

This paper described a simulated Virtual Prototyping test bed to evaluate the feasibility of deploying an Augmented Reality assisted non-contact building damage reconnaissance method in the field. The research demonstrated the effectiveness of VR-assisted Virtual Prototyping in evaluating and demonstrating new reconnaissance methods where full-scale physical test bed experimentation is impractical. The experimental plan constructed a ten-story graphical building capable of expressing the internal structural damage through the texture displacement on the surface. LSD can detect the shifted building edge on the captured building image, and the final corner coordinate is triangulated through the intersections between the detected vertical building edge and the projected horizontal baseline. The experiment results with ground true location and orientation data are satisfactory for damage detection requirements. The results also highlight the conditions for achieving the ideal measurement accuracy, for example observing distance, angle, and image resolution. The experimental results with instrumental errors reveal the bottleneck for the proposed method in the field implementation conditions. While the state of the art RTK-GPS can meet the location accuracy requirement, the electronic compass is not accurate enough to supply qualified measurement data, suggesting that alternative survey-grade orientation measurement methods must be identified to replace electronic compasses. The conducted sensitivity analysis developed a clear matrix revealing the relation between instrument accuracy and accuracy of computed drift, so the proposed method's practical implementation can evolve with choices made for higher accuracy instruments than the ones tested. The authors acknowledge that the sensitivity matrix developed from the virtual prototyping may have limitations and needs to be further validated in a real environment setting. For example, the dynamic illumination may bring challenges to the edge detection. Furthermore the estimation method assumes ground true geometric building information is available. It is possible that in reality, such information contains uncertainty or is possibly unavailable for older buildings. The current virtual prototyping has not modeled such data uncertainty. The open source code for the virtual prototyping and its sensitivity analysis is available at <http://pathfinder.engin.umich.edu/software.htm>.