Intelligent Analysis of Hydraulic Concrete Vibration Time Based on Convolutional Neural Network

Abstract

The vibrating of concrete is one of the most important procedures that directly determines the quality of construction projects. The concrete vibration quality in field construction is mainly judged by the experience of workers, lacking quantitative indicators, and necessary supervision. However, the lack of research about concrete vibration quality led to these problems are still existing. There are some methods are proposed that are too difficult, or too expensive to use in field construction. Combined with the pouring project of Jianquan Pumped Storage Power Station in Yunyang, China, this research developed an intelligent detection system for concrete vibration time. The system took the convolutional neural network as the basic framework, and divided the concrete vibration process into three different states: vibrating, not vibrating, and no vibration tube, realized the concrete vibration time through the analysis of concrete vibration video data. The detection of concrete vibration process video with multiple stages in the actual project shows that the detection error of the system for each state is kept within 1 s, the accuracy is high, which can meet the quality management requirements of the construction process. The system can be quickly deployed to the construction site by using mobile phones, cameras, and other common equipment. It has the advantages of simple frame structure, low hardware requirements, and accurate detection results. In addition, the current system training process is only for the concrete pouring process of Jianquan Power Station, and the training sample can be further expanded in the future to enhance the applicability and accuracy of the system in other engineering applications in order to play a better role.