Abstract: Mining subsidence prediction is an important part of the study of mining subsidence,most of the existing mining subsidence prediction softwares are developed based on the rectangle subdivision model,which have the disadvantages of poor adaptability and low accuracy of the mining subsidence prediction results.Aiming at the defects of the existing mining subsidence softwares,firstly,based on the triangle partition model and the corrected angular point coordinates,the mining working face is divided into multiple independent triangles based on VB programme language;then,the double variable step size Simpson algorithm is adopted to calculate the displacement and deformation caused by each triangle mining working face based on probability integral method;finally,the mining subsidence prediction results of each triangle mining working face are conducted superposition to realize the mining subsidence prediction of each mining working face with high precision and the visual output of the mining subsidence prediction results.The experimental results show that:①the mining subsidence prediction method based on triangle subdivision algorithm overcomes the shortcomings of the rectangular subdivision such as processing the mining working face with high difficult,equivalent subdivision and so on,the engineering needs can be satisfied,at the same time,the mining subsidence monitoring data processing flow is simplified,therefore,the accuracy and reliability of the mining subsidence prediction results are improved;②the mining subsidence prediction and parameters inversion system development based on triangle subdivision algorithm with prefect stability and practicability,it can provide scientific basis for the design of "three-underground"coal mining scheme,environmental protection of mining area and land reclamation of mining area,besides that,it also contribute to the realization of green mining.

Key words: Mining subsidence, Triangle subdivision, Rectangle subdivision, Probability integral method, Simpson algorithm, Green mining, VB