Coal mining-induced surface subsidence is a complex

multi-dimensional dynamic process. Dynamic prediction plays a crucial role in determining the magnitude of deformation at any location and time above the goaf

thereby safeguarding ground infrastructure and human life. To achieve dynamic prediction of surface subsidence

an improved Weibull time function model with a single model parameter was developed to address the limitations of the complex structure of the existing model

based on mining subsidence theory. This enhanced model accurately describes surface point subsidence

subsidence velocity

and subsidence acceleration. Additionally

a new dynamic prediction model for mining subsidence was established by integrating the improved Weibull time function model with the existing surface subsidence basin model. The method for determining model parameters and their impact on the shape of the subsidence basin was elaborated in detail. The model’s prediction accuracy and applicability were validated using measured data from working faces 313 and 3214 in a mine. The results indicate that the maximum root mean square error of the improved Weibull time function model is 52 mm

with a maximum relative error of 2.1%

representing a 60.3% and 64.4% improvement

respectively

over the previous version. The predicted subsidence curve shape of the mining subsidence dynamic prediction model aligns with the measured subsidence curve shape

with a maximum root mean square error of 17 mm and a maximum relative error of 1.76%

demonstrating the model’s ability to predict surface subsidence processes with high applicability and reliability. These research findings offer valuable insights for the dynamic prediction of surface subsidence in mining areas.