Safety Prediction Method of Urban Construction FP Based on Multi-Order Spatio-Temporal Features

With the process of urbanization, the scale of urban foundation pit construction is also increasing. To increase the accuracy of safety prediction for foundation pit construction and increase the safety performance of urban construction, BP is applied, and genetic algorithm is used to optimize it to improve prediction. Meanwhile, through the time dimension and space dimension, the method based on multi-order spatio-temporal features is used to accurately predict the modeling displacement. The maximum error of horizontal displacement prediction for multi-order spatio-temporal features is 0.21, which is 34.37% lower than the maximum error of single-order prediction 0.32. The maximum error of vertical displacement prediction for multi-order spatio-temporal features is 0.05, which is 68.75% lower than the maximum error of single-order prediction 0.16. In summary, the proposed method effectively improves the safety monitoring accuracy of urban foundation pit construction projects, reduces the incidence of safety accidents, and provides a guarantee for the stable and safe development of cities.