In this paper, a series of numerical experiments are carried out on the anti-scour device combined with perforated baffle and ring-wing plate. In addition, the optimal dimensions and location of the combined device are obtained: The perforated ratio of the baffle is S = 20%, the distance from the center of the bridge pier is L = 2d (d is the diameter of the bridge pier), and the ring-wing plate is located at H = 1/3h (h is the water depth). To verify the effect of the anti-scour device, the scour characteristics and flow field are further investigated. Compared with single pier and single ring-wing plate, the results revealed that the combined device with the optimal dimensions is of great anti-scour performance. Moreover, the maximum scour depth at the front and side of the pier reduced by 84.20% and 78.95%, which is better than the single ring-wing plate and other combined conditions in the orthogonal experiments. Due to the diversion of perforated baffle and ring-wing plate, the flow velocity at the pier side near the bed surface decreases by 30.7%, and the down-flow is almost eliminated on the vertical plane. Furthermore, the turbulent kinetic energy at different horizontal and vertical planes is reduced due to the reduction in horseshoe vortex and wake flow. Based on the investigation presented herein, the combined device is a promising tool for mitigating scour around the bridge pier.

For the maintenance of weathering steel structure facilities, it is necessary to evaluate the corrosion grade of the rust layer on the surface regularly. At present, the corrosion grade classification of weathering steel is mainly based on the human-eye inspection. In this paper, a deep learning method using a convolutional neural network for evaluating the corrosion grade of weathering steel is proposed to save time and manpower. Firstly, the image dataset of the corrosion steel plate was established using salt spray tests. Then, a CNN architecture named VGG-Corrosion was designed to evaluate the corrosion grade of the corroded steel plate. The effect of the learning rate, transfer learning, and batch size was also investigated to clarify the best hyperparameter configurations to train a powerful corrosion grade classification model. Under the best combination of considered hyperparameters, the mean average accuracy for the corrosion grade evaluation of the test results is 90.96%. The testing results indicated that the CNN based corrosion grade recognition for weath-ering steel plate is prospective, which would be helpful for safety evaluation of steel structures.