ارائه یک روش ابتکاری برای تشخیص آسیب در تیرهای عرشه پل‌های فولادی و بتنی با ترکیب پاسخ‌های جابه‌جایی-زمان حسگرها و تبدیل موجک گسسته (DWT) و ارائه روشی جدید برای تعیین مؤثرترین مکان‌های نصب حسگرها.

An innovative bridge deck girders damage detection approach by combining sensor displacement-time responses and Discrete Wavelet Transform (DWT), and proposing a novel method to determine the most effective sensor placement locations.



Abstract



Over time, bridge structural members, particularly their steel and concrete girders, experience a reduction in stiffness, fatigue cracking, and localized damage due to environmental factors (such as temperature fluctuations, humidity, and corrosion), dynamic loads from passing vehicles, and potential design or construction flaws. A fundamental challenge in bridge structural health monitoring (SHM) is the accurate and timely detection of damage under actual operating conditions and moving loads. This study introduces an innovative approach combining sensor displacement-time responses with the Discrete Wavelet Transform (DWT) to detect damage and pinpoint its exact location; furthermore, a novel method is proposed to determine the most effective sensor placement, aiming to minimize the total number of sensors required.



This study introduces a novel approach that identifies damage and pinpoints its exact location by combining sensor displacement-time responses with the Discrete Wavelet Transform; it further proposes a new method to determine optimal sensor placement aimed at minimizing the number of sensors required. Based on the extraction of a "normalized wavelet energy difference" index—derived from comparing structural responses in healthy and damaged states under a moving load—this approach ultimately provides a practical, implementable solution for effective instrumentation in structural health monitoring projects.



To this end, two bridges—the Mill Pond Powder bridge (steel girder) in the USA and the Abbas-Abad bridge (concrete girder) in Iran (designed by the author)—were modeled using CSI Bridge software; the validity of these models was confirmed by comparing results with previous studies and field load test data from similar systems, respectively. To realistically simulate damage, fatigue cracks were modeled by reducing element stiffness at critical locations—identified through stress analysis under AASHTO and Publication 139 code loads—across three critical scenarios (including support regions and mid-span). The deck's displacement-time response under a constant-speed moving load was extracted using a pseudo-dynamic method and subsequently analyzed via the Discrete Wavelet Transform (DWT) using a custom code developed in MATLAB.



The algorithm developed in this environment, featuring automated analysis capabilities, evaluates various defined mother wavelets and decomposition levels, ranking them from highest to lowest based on the "normalized energy difference" between the healthy and damaged states. In this process, maximizing this difference serves as the primary criterion for selecting the optimal configuration, as higher values indicate greater damage clarity and a superior ability to distinguish responses, thereby yielding the most accurate results. The findings demonstrated that the precise location of the damage can be identified by ranking the "wavelet normalized energy difference" values and analyzing the timing of the index's peak relative to the load speed.



Furthermore, the analysis of the behavior at various points along the deck yielded a key insight regarding sensor placement: based on Saint-Venant's principle and beam theory, the beam web—located near the site of a flange crack—reveals the damage more clearly due to drastic changes in stress distribution, whereas at greater distances, the flange on the same side as the damage provides a more discernible response. This approach has led to the development of a comprehensive, operational flowchart for the health monitoring of actual bridges using a limited number of sensors.