The ability to perform pixel-wise segmentation on pavement cracks in real-time is paramount in road service condition assessment and maintenance decision-making practices. Recent deep learning detection models are focused on detection accuracy and require a large number of computing sources and long run times. However, highly efficient and accelerated models with acceptable accuracy in real-time pavement crack detection tasks are required but hard to achieve. In this work, we present a customized deep learning model architecture named Efficient Crack Segmentation Neural Network (ECSNet) for accelerated real-time pavement crack detection and segmentation without compromising performance. We introduce some novel parts, including small kernel convolutional layers and parallel max pooling and convolutional operation, into the architecture for crack information quickly extraction and model’s parameter reduction. We test latency and accuracy trade-offs of our proposed model using the DeepCrack Dataset. The results demonstrate strong performance in both accuracy and efficiency compared to other state-of-the-art models including DeepLabV3, FCN, LRASPP, Enet, Unet and DeepCrack. It is promising that ECSNet obtains the second place with an F $_{\mathbf{1}}$ score of (84.45%) and an Intersection over Union (IoU) of 73.08%. Furthermore, our model gains the largest Frames Per Second (FPS) and lowest training time among all the models which is 73.29 and 5011 seconds, respectively. It maintains a good balance between accuracy and efficiency metrics.

The regional multi-hazards risk assessment poses difficulties due to data access challenges, and the potential interactions between multi-hazards and social vulnerability. For better natural hazards risk perception and preparedness, it is important to study the nature-hazards risk distribution in different areas, specifically a major priority in the areas of high hazards level and social vulnerability. We propose a multi-hazards risk assessment method which considers social vulnerability into the analyzing and utilize machine learning-enabled models to solve this issue. The proposed methodology integrates three aspects as follows: (1) characterization and mapping of multi-hazards (Flooding, Wildfires, and Seismic) using five machine learning methods including Naïve Bayes (NB), K-Nearest Neighbors (KNN), Logistic Regression (LR), Random Forest (RF), and K-Means (KM); (2) evaluation of social vulnerability with a composite index tailored for the case-study area and using machine learning models for classification; (3) risk-based quantification of spatial interaction mechanisms between multi-hazards and social vulnerability. The results indicate that RF model performs best in both hazard-related and social vulnerability datasets. The most cities at multi-hazards risk account for 34.12% of total studied cities (covering 20.80% land). Additionally, high multi-hazards level and socially vulnerable cities account for 15.88% (covering 4.92% land). This study generates a multi-hazards risk map which show a wide variety of spatial patterns and a corresponding understanding of where regional high hazards potential and vulnerable areas are. It emphasizes an urgent need to implement information-based prioritization when natural hazards coming, and effective policy measures for reducing natural-hazards risks in future.

Cement hydration kinetics, characterized by heat generation in early-age concrete, poses a modeling challenge. This work proposes a physics-informed neural network (PINN) named PINN-CHK designed for cement hydration kinetics, to predict early-age temperature rises in cement paste. PINN-CHK leverages data-driven solutions to craft a high-fidelity prediction model, encompassing material properties and maturity functions in cement hydration. Trained on heated cement paste data, it simultaneously fits experimental results and underlying physics, yielding a mesh-free simulation. Incorporating governing partial differential equations (PDEs), and initial and boundary conditions into its loss function, PINN-CHK architecture undergoes rigorous benchmark testing, demonstrating unparalleled predictive accuracy compared to conventional deep-learning methods. It excels in predicting complete temperature fields during spatial–temporal cement hydration, achieving a remarkable relative L2 error as low as 0.00341. PINN-CHK achieves exceptional convergence and accuracy with only 5% of the training data, ushering in a new era in this crucial field. This innovative approach bridges the gap between theory and practice, offering an attractive alternative to conventional finite element solvers for enhanced comprehension of cement hydration kinetics and concrete maturity and strength development in cement-based materials.

Pavement crack segmentation using deep learning methods can improve crack segmentation accuracy, but in many cases the training dataset is lacking or uneven, making it insufficient to train an accurate segmentation model. In this work, an integrated APC-GAN and AttuNet framework is proposed as an automated pavement surface crack pixel-level segmentation solution for small training datasets. First, an Automated Pavement Crack Generative Adversarial Network (APC-GAN) is designed for the pavement cracks data as an image augmentation method, which is modified and improved from a traditional Deep Convolutional Generative Adversarial Network (DCGAN). Then, a novel pixel-level semantic segmentation structure, Attention modified U-Net (AttuNet), is proposed by introducing the attention module into the convolutional network structure. In order to assess the performance of our proposed framework, an open-source dataset DeepCrack is used, which only contains 300 training images. The results show that our proposed APC-GAN could augment the datasets by producing more clear and distinct pavement images than DCGAN and the generated images could feature more diversity than traditional image augmentation methods. APC-GAN demonstrated higher accuracy than DCGAN and traditional image augmentation methods. Apparently, our proposed APC-GAN and AttuNet framework gains the highest value in the evaluation metrices, including recall, F1 score, mean Intersection over Union (mIoU) and mean Pixel Accuracy (mPA) among all models including U-Net, DeepLabv3, FCN, and LRASPP.

Many studies were conducted to find possible strategies for reducing the urban heat island (UHI) effect during the hot summer months. One of the largest contributors to UHI is the role that paved surfaces play in the warming of urban areas. Solar-reflective cool pavements stay cooler in the sun than traditional pavements. Pavement reflectance can be enhanced by using a reflective surface coating. The use of heat-reflective coatings to combat the effects of pavements on UHI was pre-viously studied but no consistent conclusions were drawn. To find a conclusive solution, this work focuses on the abilities of heat-reflective pavement coatings to reduce UHI in varying weather conditions. Within this context, both concrete and asphalt samples were subject to a series of per-formance tests when applied to a heat-reflective coating, under the influence of normal, windy, and humid conditions. During these tests, the samples were heated with a halogen lamp and the surface temperature profile was measured using an infrared thermal camera. The air temperature was recorded with a thermometer, and the body temperature at multiple depths of the samples was measured using thermocouples. The results from these tests show that the effectiveness of the heat-reflective coating varies under different weather conditions. For instance, the coated samples were about 1 °C cooler for concrete and nearly 5 °C cooler for asphalt, on average. However, this temperature difference was reduced significantly under windy conditions. As such, the findings from this work conclude that the heat-reflective coatings can effectively cool down the pavement by increasing the surface albedo, and thus might be a viable solution to mitigate UHI impacts in the city/urban areas.

Dynamic response of pavement provides service condition information and helps with damage prediction, while limited research is available with the simulation of pavement vibration response for evaluating roadway service condition. This paper presents a numerical model for the analysis of the pavement vibration due to the dynamic load created by a passing vehicle. A quarter vehicle model was used for the determination of the vehicle moving load. Both random and spatial characteristics of the load were considered. The random nonuniform moving load was then introduced in a 3D finite element model for the determination of the traffic‐induced pavement vibration. The validated numerical model was used to assess the effects of dynamic load, material properties, and pavement structures on pavement vibration response. Numerical analyses showed that the vibration modes changed considerably for the different roadway service conditions. The vibration signals reflect the level of road roughness, the stiffness of the pavement materials, and the integrity of pavement structure. The acceleration extrema, the time‐domain signal waveform, the frequency distribution, and the sum of squares of Fourier amplitude can be potential indexes for evaluating roadway service condition. This provides recommendations for the application of pavement vibration response in early‐warning and timely maintenance of road.