Simulation And Prediction Of the Safety Risk Of Tower Crane For Super High-rise Buildings Through Back Propagation Neural Network

Abstract

High-rise buildings especially super high-rise buildings are booming with the rapid urbanization because the urban population increases dramatically. For the sustainable development of super high-rise buildings, construction safety is an important issue. Since tower cranes are widely applied on the construction of super high-rise buildings as the lifting and transportation equipment, their safety is closely related to the safety of construction sites. In this study, a reasonable safety evaluation system with clear input and output indicators was first established. Then, the input index was quantified, and their weights were determined according to their importance which were ranked based on the fourth-class division method of safety grade. Subsequently, the neural network model was developed. 90 tower crane accidents were evaluated quantitatively which provided the data to train and verify the model. The results show that prediction of neural network model is reliable. The trained neural network model is capable to predict the tower crane risks of high-rise buildings, which can be helpful to the construction safety management.