Finite-element analysis (FEA) for structures has been broadly used to conduct stress analysis of various civil and mechanical engineering structures. Conventional methods, such as FEA, provide high fidelity results but require the solution of large linear systems that can be computationally intensive. Instead, Deep Learning (DL) techniques can generate results significantly faster than conventional run-time analysis. This can prove extremely valuable in real-time structural assessment applications. Our proposed method uses deep neural networks in the form of convolutional neural networks (CNN) to bypass the FEA and predict high-resolution stress distributions on loaded steel plates with variable loading and boundary conditions. The CNN was designed and trained to use the geometry, boundary conditions, and load as input to predict the stress contours. The proposed technique’s performance was compared to finite-element simulations using a partial differential equation (PDE) solver. The trained DL model can predict the stress distributions with a mean absolute error of 0.9% and an absolute peak error of 0.46% for the von Mises stress distribution. This study shows the feasibility and potential of using DL techniques to bypass FEA for stress analysis applications.
It is of great significance to quickly detect underwater cracks as they can seriously threaten the safety of underwater structures. Research to date has mainly focused on the detection of above-water-level cracks and hasn’t considered the large scale cracks. In this paper, a large-scale underwater crack examination method is proposed based on image stitching and segmentation. In addition, a purpose of this paper is to design a new convolution method to segment underwater images. An improved As-Projective-As-Possible (APAP) algorithm was designed to extract and stitch keyframes from videos. The graph convolutional neural network (GCN) was used to segment the stitched image. The GCN’s m-IOU is 24.02% higher than Fully convolutional networks (FCN), proving that GCN has great potential of application in image segmentation and underwater image processing. The result shows that the improved APAP algorithm and GCN can adapt to complex underwater environments and perform well in different study areas.
This paper proposes a framework for critical element identification and demolition planning of frame structures. Innovative quantitative indices considering the severity of the ultimate collapse scenario are proposed using reinforcement learning and graph embedding. The action is defined as removing an element, and the state is described by integrating the joint and element features into a comprehensive feature vector for each element. By establishing the policy network, the agent outputs the Q value for each action after observing the state. Through numerical examples, it is confirmed that the trained agent can provide an accurate estimation of the Q values, and handle problems with different action spaces owing to utilization of graph embedding. Besides, different behaviors can be learned by varying hyperparameters in the reward function. By comparing the proposed method and the conventional sensitivity index-based methods, it is demonstrated that the computational cost is considerably reduced because the reinforcement learning model is trained offline. Besides, it is proved that the Q values produced by the reinforcement learning agent can make up for the deficiencies of existing indices, and can be directly used as the quantitative index for the decision-making for determining the most expected collapse scenario, i.e., the sequence of element removals.
With increasing population and limitation of availability of land, tall buildings supported on piled raft foundations are increasingly used in the modern world. To increase the ratio of floor area to height, and to fulfill storage and parking facilities requirements, these tall buildings usually have more than one basement level. Conventionally, during the foundation design, engineers have not considered the basement wall contribution to resisting lateral load induced by earthquake or wind and this can result in an uneconomical construction of foundations. In this research work, an experimental study was performed on small-scale models, in order to study basement wall contribution, and the raft contribution including for piled raft foundations, to resisting lateral load. Three configurations of piles in 2 × 2, 2 × 3, and 3 × 3 patterns were tested as a pile group, piled raft and piled raft with a basement wall. Results show that when a basement wall is present, the lateral displacement decreases and the demand on each pile decreases. The piled raft design can become more economical for tall buildings if the basement’s walls are taken into account.
Engineered cementitious composites (ECC), also known as bendable concrete, were developed based on engineering the interactions between fibers and cementitious matrix. The orientation of fibers, in this regard, is one of the major factors influencing the ductile behavior of this material. In this study, fiber orientation distributions in ECC beams influenced by different casting techniques are evaluated via numerical modeling of the casting process. Two casting directions and two casting positions of the funnel outlet with beam specimens are modeled using a particle-based smoothed particle hydrodynamics (SPH) method. In this SPH approach, fresh mortar and fiber are discretized by separated mortar and fiber particles, which smoothly interact in the computational domain of SPH. The movement of fiber particles is monitored during the casting simulation. Then, the fiber orientations at different sections of specimens are determined after the fresh ECC stops flowing in the formwork. The simulation results show a significant impact of the casting direction on fiber orientation distributions along the longitudinal wall of beams, which eventually influence the flexural strength of beams. In addition, casting positions show negligible influences on the orientation distribution of fibers in the short ECC beam, except under the pouring position.
This paper utilized granulated blast furnace slag (GBFS), fly ash (FA), and zeolite powder (ZP) as the binders of ternary geopolymer concrete (TGC) activated with sodium silicate solution. The effects of alkali content (AC) and alkaline activator modulus (AAM) on the compressive strength, flexural tensile strength and elastic modulus of TGC were tested and the SEM micrographs were investigated. The experimental results were then compared with the predictions based on models of mechanical properties, and the amended models of TGC were proposed taking account of the effects of AC and AAM. The results indicated that increasing AC and reducing AAM which were in the specific ranges (5% to 7% and 1.1 to 1.5, respectively) had positive effects on the mechanical properties of TGC. In addition, the flexural tensile strength of TGC was 27.7% higher than that of OPC at the same compressive strength, while the elastic modulus of TGC was 25.8% lower than that of OPC. Appropriate prediction models with the R2 of 0.945 and 0.987 for predicting flexural tensile strength and elastic modulus using compressive strength, respectively, were proposed. Fitting models, considering the effects of AC and AAM, were also proposed to predict the mechanical properties of TGC.
Construction industries have started to utilize manufactured sand (MS) as an effective alternative for river sand in concrete. High-grade parent rocks are crushed to obtain MS, which also produces a considerable amount of microfine aggregate (MFA). The higher percentage of MFA could lead to both positive and negative effects on the performance of cement-based mixes. This research was done to examine the influence of varying MFA levels, specifically 0%, 3%, 6%, 9%, and 12% (by weight) as the partial replacements of MS on bleeding and plastic shrinkage cracking of concrete. In addition to the varying MFA levels, some concrete mixes also included fly ash (FA) and superplasticizer to investigate the effect of free-water content in the mixes. The bleeding test data were taken as on-site measurements, while the cracks from the plastic shrinkage cracking test were evaluated using an image processing technique. The results concluded that the MFA replacements and the effective water-to-cement ratio have a significant effect on the selected concrete properties. With the increasing replacement levels, cumulative bleeding and crack initiation life gradually decreased, while a progressive increase was observed for crack width, crack length, and crack area.
Emulsified asphalt is the primary material for preventive maintenance and cold-mix paving, but its low cohesive strength and poor mechanical properties limit its wide application, even with polymer modification. In this study, Styrene-Butadiene Rubber (SBR) emulsified asphalt was modified with nano-cellulose materials, namely nano paper-cellulose (NPC) and wood-derived nano-cellulose (WDC), to improve its properties. A novel preparation method of nano-cellulose solution was developed, including blending, ultrasonic stirring, and centrifugal treatment. Four types of nano-cellulose solution (0.5% NPC, 0.5%, 1.0%, and 1.5% WDC by weight of water) were selected. The microscopy analysis indicated that 0.5% WDC emulsion had a smaller particle size than 1.5% WDC emulsion. The rheology test indicated that WDC modified residue improved rutting resistance with the increased solution dosage due to the cross-linking effect, but its creep-and-recovery performance was worse than that of SBR emulsion residue. The NPC modified binder had a higher rutting factor than WDC modified binder at the same dosage after short-term aging. In addition, 1.0% WDC could be regarded as the optimal dosage in terms of fatigue and low-temperature performance. Furthermore, Fourier Transform Infrared Spectroscopy (FTIR) results showed that 0.5% NPC modified residue performed better in long-term aging resistance compared with 0.5%WDC modified asphalt.
The International Energy Agency (IEA) states that global energy consumption will increase by 53% by 2030. Turkey has 70% of the world’s perlite reserves, and in order to reduce energy consumption a thermal insulation panel was developed in Turkey using different particle sizes of expanded perlite (EP). In this study, 0–1.18 mm (powder) and 0–3 mm (granular) EP particle sizes were selected, since they have the lowest thermal conductivity coefficients among all the particle sizes. In addition, an alkali activator solution was used as a binder in the mixtures. The alkaline activator solution was obtained by mixing sodium hydroxide solution (6, 8, 10, and 12 mol·L−1) and sodium silicate (Module 3) at the different ratios of Na2SiO3 to NaOH of 1, 1.5, 2, and 2.5. This study aimed to experimentally determine the optimum binder and distribution ratio of EP, with the lowest coefficient of thermal conductivity and the lowest density. The lowest thermal conductivity and the lowest density were determined as 0.04919 W·m−1·K−1 and 133.267 kg/m3, respectively, in the sample prepared with 83.33% powder-size EP, 6 mol·L−1 sodium hydroxide solution, and ratio of Na2SiO3 to NaOH of 1.5. The density, thermal conductivity, and compressive strength of the sample showed the same trends of behavior when the Na2SiO3 to NaOH ratio was increased. In addition, the highest compressive strength was measured in 12 mol·L−1 NaOH concentration regardless of particle size. In conclusion, the study predicts that the EP-based thermal insulation panel can be used as an insulation material in the construction industry according to the TS825 Thermal Insulation Standard.
