Herein, two novel Physics Informed Neural Network (PINN) architectures are proposed for output-only system identification and input estimation of dynamic systems. Using merely sparse output-only measurements, the proposed PINNs architectures furnish a novel approach to input, state, and parameter estimation of linear and nonlinear systems with multiple degrees of freedom. These architectures are comprised of parallel and sequential PINNs that act upon a set of ordinary differential equations (ODEs) obtained from spatial discretization of the partial differential equation (PDE). The performance of this framework for dynamic system identification and input estimation was ascertained by extensive numerical experiments on linear and nonlinear systems. The advantage of the proposed approach, when compared with system identification, lies in its computational efficiency. When compared with traditional Artificial Neural Networks (ANNs), this approach requires substantially smaller training data and does not suffer from generalizability issues. In this regard, the states, inputs, and parameters of dynamic state-space equations of motion were estimated using simulated experiments with “noisy” data. The proposed framework for PINN showed excellent great generalizability for various types of applications. Furthermore, it was found that the proposed architectures significantly outperformed ANNs in generalizability and estimation accuracy.

Transfer learning techniques for structural health monitoring in bridge-type structures are investigated, focusing on model generalizability and domain adaptation challenges. Finite element models of bridge-type structures with varying geometry were simulated using the OpenSeesPy platform. Different levels of damage states were introduced at the midspans of these models, and Gaussian-based load time histories were applied at mid-span for dynamic time-history analysis to calculate acceleration data. Then, this acceleration time-history series was transformed into grayscale images, serving as inputs for a Convolutional Neural Network developed to detect and classify structural damage states. Initially, it was trained and tested on datasets derived from a Single-Source Domain structure, achieving perfect accuracy (1.0) in a ten-label multi-class classification task. However, this accuracy significantly decreased when the model was sequentially tested on structures with different geometry without retraining. To address this challenge, it is proposed that transfer learning be employed via feature extraction and joint training. The model showed a reduction in accuracy percentage when adapting from a Single-Source Domain to Multiple-Target Domains, revealing potential issues with non-homogeneous data distribution and catastrophic forgetting. Conversely, joint training, which involves training on all datasets except the specific Target Domain, generated a generalized network that effectively mitigated these issues and maintained high accuracy in predicting unseen class labels. This study highlights the integration of simulation data into the Deep Learning-based SHM framework, demonstrating that a generalized model created via Joint Learning utilizing FEM can potentially reduce the consequences of modeling errors and operational uncertainties unavoidable in real-world applications.

A vast majority of existing substandard reinforced concrete (RC)-framed buildings are prone to seismic damage because of improper design and construction applications as well as poor material properties. In this study, field investigation and analytical research have been conducted for adjacent buildings, which were exposed to an earthquake with a magnitude of 5.0 occurred in Çiçekdağı/Kırşehir/Turkey on January 10, 2016. A building with a basement story was severely damaged following the earthquake and collapsed eight days after the main shock whereas the other building with no basement did not experience significant damage and survived. Except for the presence of the basement in the collapsed building, both buildings have the same structural system and material properties. To let in daylight, ribbon windows were constructed on the partial infill walls adjacent to some of the columns at the basement story level, which resulted in the short-column phenomenon in the collapsed building. Material tests were conducted on the concrete and reinforcing steel samples taken from the collapsed building to determine their mechanical properties. Then, analytical models of both buildings with masonry infill walls were generated based on the information obtained from the site observations, the blueprints of the buildings and the results of material tests. Nonlinear static pushover analyses were performed to obtain the capacity curves of the buildings. Analyses results showed that the formation of short-column damage, which was visually observed in the field, in the basement because of imposed excessive shear demand was found to be the main reason for the total collapse as well as the high axial load level of columns.

Shape memory alloys (SMAs) have been utilized as an alternative to conventional materials for seismic retrofit applications in the last two decades. The effectiveness of SMAs to provide further improvement in seismic behavior of structures was demonstrated satisfactorily. Three 2/3-scaled, one-bay and one-storey substandard RC frames were constructed to represent the seismically vulnerable buildings with several deficiencies. The first RC frame was tested as a reference specimen under quasi-static reversed cyclic loading, which caused flexural plastic hinges at the column ends. Then, the rest of the two RC frames were upgraded with superelastic copper–aluminum–manganese (CuAlMn) alloy bars and conventional steel bars with the aim of enhancing seismic performance of substandard RC frames. The upgrading materials were attached to the RC frames through a retrofitting mechanism to provide tension-only retrofitting bars. Hence, the RC frame retrofitted by conventional steel bars was exposed to residual displacements after they yielded. The SMA-upgraded frame showed flag shaped hysteresis curve due to its superelastic behavior and caused considerable reduction in the residual displacement of RC frames.