Research on Intelligent Dynamic Prediction of Deep Foundation Pit Construction Schedule under Uncertain Background

With the rapid economic development of China, there has been an increase in the number of development projects involving urban underground spaces. As a result, deep foundation pit projects have become an inevitable construction activity. Compared to other construction projects, deep foundation pit projects exhibit considerable uncertainty, which can arise during their design, construction, and management stages. Progress is one of the three major indicators of project management. However, due to the significant changes in construction conditions over time and space during the excavation process, accurately forecasting the progress of deep foundation pit projects is extremely challenging. Relying solely on the planned progress before construction is insufficient to provide an accurate prediction of their progress. Therefore, this paper combines computer vision technology, discrete event simulation, and Bayesian update theory to achieve intelligent and dynamic prediction of deep foundation pit construction progress. In this study, we first employ computer vision algorithms to identify the status and quantity of dump trucks at the construction site by analyzing construction videos, allowing us to estimate the excavation efficiency of the earthwork. Next, we decompose the deep foundation pit construction process and establish a discrete event simulation model. By coupling Bayesian update theory with the discrete event simulation model, we input the data collected through intelligent means based on computer vision into the established model. Subsequently, we run the model to realize dynamic prediction of deep foundation pit construction progress. Finally, we apply the proposed method to a practical engineering case to validate its effectiveness, which significantly enhances the intelligence and accuracy of deep foundation pit construction progress prediction.