Structural health monitoring (SHM) has been widely employed in civil infrastructures for a number of years. Real‐time monitoring of civil projects involves the utilization of diverse sensors. Nevertheless, accurately assessing the actual condition of a structure can pose challenges due to the existence of anomalies in the collected data. Abnormalities in this context often arise from a variety of factors, including extreme weather conditions, malfunctioning sensors, and structural impairments. The existing condition of anomaly detection is significantly impeded by this disparity. Online detection of anomalies in SHM data plays a crucial role in promptly assessing the status of structures and making informed decisions. In vibration‐based SHM, enhanced frequency domain decomposition (EFDD) is one of the most used methods in the frequency domain. The signal output obtained from EFDD also includes the frequencies of the structures, which is a holistic evaluation. The findings of frequency measurements are influenced by the presence of structural damages. Extracting damage‐sensitive characteristics from structural response has emerged as a complex task. Deep learning approaches have garnered growing interest due to their capacity to efficiently extract high‐level abstract features from raw data. Within the scope of the study, a novel approach based on anomaly detection of changes in the signal output obtained using the EFDD was developed with autoencoders in deep learning. The performance of the novel approach was examined depending on different noise ratios (0%, 0.5%, 1%, 1.5%, and 2.0%) using the Z24 Bridge dataset. In the autoencoder training model, an autoencoder model containing a 4 Conv1D layer encoder–decoder as 128 × 64 × 64 × 128 was designed. By using the signal data of the first singular values obtained with the EFDD method, grouping was made with the labels “training data (1260 pieces),” “undamaged new data (250 pieces),” and “damaged new data (320 pieces).” In addition, the upper limit of the reconstruction error was calculated as 810 using the training data in the autoencoder model. The filtered reconstruction error values obtained were compared under different noise levels. At the end of the study, it was concluded that the novel approach works effectively under different noises and can be used in anomaly detection.

The use of timber as a construction material is based on much older dates than concrete and steel. Therefore, the preservation of historical timber structures such as bridges, monuments, towers, mosques, etc. is very important for cultural heritage. The perspective of preservation for this kind of structures has been developed and has gained more importance in the last two decades. However, there are some issues that need to be addressed. Timber is an anisotropic material and is not fire resistant. Also, moisture causes significant swelling. So, structural behavior should be carefully evaluated by experimental techniques before and after restoration and verified numerically using finite element models. This article presents a detailed study on the structural condition assessment of a restored historical timber mosque: its finite element analysis (FEA), ambient vibration tests, model updating to minimize the differences and reflect the current situation, and dynamic analyses before and after updating procedure. The Kuşluca historical timber mosque located in the Sürmene District of Trabzon, Turkey was selected as an application. The mosque was built in the early 20th century and includes two floors. Significant timber decorative elements are available in the entrance door, minbar, and mihrab. The restoration projects began in 2008, and the restored mosque was opened in 2011. Finite element modeling of the mosque was accomplished using SAP2000 software (Computers and Structures, Inc., Walnut Creek, California), considering the restoration project drawings. Modal analysis was performed using orthotropic material properties, considering literature review, to determine the initial dynamic characteristics. Nondestructive experimental measurements were conducted after construction to validate the numerical results using in situ testing. Ambient vibration-based system identification was employed using the Enhanced Frequency Domain Decomposition method in the frequency domain and Stochastic Subspace Identification method in the time domain. The first three natural frequencies were obtained between 5.160 Hz to 7.153 Hz and 3.960 Hz to 5.873 Hz numerically and experimentally, respectively. There is close agreement between mode shapes, but 30.0 % differences in natural frequencies. To minimize the differences, the finite element model of the timber mosque is updated using the manual model updating procedures, with a changing of material properties to reflect the real structural behavior. The maximum differences are reduced to below the acceptable limits as 5 %. To evaluate the structural behavior and determine the model updating effect, linear dynamic-time history analyses are performed, and displacements with internal forces are compared to each other.

Historical timber bridges are significant cultural structures and have a long tradition. There are many pre-1950s bridges still in use today. In recent years, great resources have been devoted to the restoration of historical timber bridges to preserve and pass them to the next generations. Because wood is an orthotropic, hygroscopic, and biodegradable material, structural behavior of timber bridges should be evaluated numerically using finite-element (FE) methods before and after restoration and systematic controlled in situ, or nondestructive experimental measurements should be made. This paper considers a historical timber bridge, using numerical modeling and modal testing to determine the structural modal parameters and FE model updating for approximation of numerical results to experimental results. As an application, the Buzlupınar historical timber bridge located in the Çayeli District of Rize, Turkey, is selected. The historical Buzlupınar Bridge was built in the middle of the nineteenth century, restored at different times for different reasons such as fire and wind effects, among others, and restored in 2014–2017. The FE model of Buzlupınar Bridge was formed using commercially available software to detect the initial modal parameters referenced on relief and restoration drawings. The orthotropic material properties are selected according to the laboratory material tests. Results of nondestructive experimental measurements, ambient vibration–based system identification was obtained using the enhanced frequency domain decomposition (EFDD) method (frequency domain) and stochastic subspace identification (SSI) method (time domain). The frequency values and mode shapes obtained by the FE model and experimental measurements are compared. According to these results, six natural frequencies are obtained between 2.0 and 20.0 Hz. Although the mode shapes are compatible with each other, there is a 34.55% difference between the frequency values. In order to reduce the differences, the FE model of the historical timber bridge is updated with a manual update method. Thanks to the model updating, the maximum differences are decreased to below 1% except the fourth mode. To evaluate the influence of FE model update, time-history analyses are applied, and internal forces and displacements are presented.