The reinforced concrete (RC) member’s shear strength estimation has been experimentally studied in most cases due to its nonlinear behavior. Many empirical equations have been derived from the experimental data; however, even those adopted in the construction codes do not thoroughly and accurately describe their shear behavior. Theoretically explained equations, on the other hand, are aligned with the experiment; however, they are complicated to use in practice. As shear behavior research is data-driven, the machine learning technique is applicable. Herein, an artificial neural network (ANN) algorithm is trained with 776 experiment results collected from available publications. The raw data is preprocessed by principal component analysis (PCA) before the application of the ANN technique. The predictions of the trained algorithm using ANN with PCA are compared to those of formulae adopted in a few existing building codes. Finally, a parametric study is conducted, and the significance of each variable to the strength of RC members is analyzed.

Artificial neural networks (ANNs) are an emerging field of research and have proven to have significant potential for use in structural engineering. In previous literature, many studies successfully utilized ANNs to analyze the structures under different loading conditions and verified the accuracy of the approach. Several studies investigated the use of ANNs to analyze the shear behavior of reinforced concrete (RC) members. However, few studies have focused on the potential use of an ANN for analysis of the torsional behavior of an RC member. Torsion is a complex problem and modeling the torsional fracture mechanism using the traditional analytical approach is problematic. Recent studies show that the nonlinear behavior of RC members under torsion can be modeled using ANNs. This paper presents a comprehensive analytical and parametric study of the torsional response of RC beams using ANNs. The ANN model was trained and validated against an experimental database of 159 RC beams reported in the literature. The results were compared with the predictions of design codes. The results show that ANNs can effectively model the torsional behavior of RC beams. The parametric study presented in this paper provides greater insight into the torsional resistance mechanism of RC beams and its characteristic parameters.

Soundproofing materials are widely used within structural components of multi-dwelling residential buildings to alleviate neighborhood noise problems. One of the critical mechanical properties for the soundproofing materials to ensure its appropriate structural and soundproofing performance is the long-term compressive deformation under the service loading conditions. The test method in the current test specifications only evaluates resilient materials for a limited period (90-day). It then extrapolates the test results using a polynomial function to predict the long-term compressive deformation. However, the extrapolation is universally applied to materials without considering the level of loads; thus, the calculated deformation may not accurately represent the actual compressive deformation of the materials. In this regard, long-term compressive deformation tests were performed on the selected soundproofing resilient materials (i.e., polystyrene, polyethylene, and ethylene-vinyl acetate). Four levels of loads were chosen to apply compressive loads up to 350 to 500 days continuously, and the deformations of the test specimens were periodically monitored. Then, three machine learning algorithms were used to predict long-term compressive deformations. The predictions based on machine learning and ISO 20392 method are compared with experimental test results, and the accuracy of machine learning algorithms and ISO 20392 method are discussed.

This paper addresses the issue of using a switching time-reversal (TR) mirror for wave energy focusing to subsurface targets. The motivation stems from applications in geophysics, hydro-geology, environmental engineering, and even in therapeutic medicine. Using TR concepts, wave-focusing is straightforward and efficient, but only under ideal conditions that are, typically, unattainable in practice. The unboundedness of the subsurface that hosts the target, the TR mirror's limited aperture, and, worse, the practical need for a switching TR mirror, where recorded Dirichlet data are time-reversed as Neumann data (switching mirror), all contribute to the deterioration of the focusing resolution at the target. Herein, the development of a data filter is discussed, which is shown to be capable of overcoming the switching mirror's shortcoming, leading to improved focusing resolution. The filter's effect is demonstrated with numerical examples.