基于CUDA并行计算的煤矿奥灰突水快速准确预测方法

Fast and Accurate Prediction Method of Ordovician Lime Water Inrush in Coal Mine Based on CUDA Parallel Computing

通过分析目前我国煤矿突水预测研究的现状,提出当前条件下我国煤矿突水预测方法难以实现实时监测突水风险问题。通过将待监测区域进行有限元剖分,根据区域内钻孔水位反演计算监测区域各节点水位,从而进行突水风险预测。结合有限元突水预测方法,将英伟达公司开发的CUDA并行计算技术应用到煤矿突水防治中,详细介绍了CUDA预测突水软件编写思路和模型。编写了CUDA监测突水程序并结合钻孔水位数据对煤矿突水点进行了实际预测,与传统CPU串行计算相比,大大提高了煤矿突水预测计算速度,对实现煤矿突水快速准确预测具有参考意义。

By analyzing the current situation of water inrush prediction research in our country’s coal mines,this paper points out that the water inrush prediction method in coal mines is difficult to achieve real-time monitoring of water inrush risk under the current conditions.The author conducts finite element analysis of the area to be monitored,and inversely calculates the water level of each node in the monitoring area according to the water level of boreholes in the area,aiming to predict the risk of water inrush.Combined with the above-mentioned finite element water inrush prediction method,the author applies CUDA parallel computing technology developed by Nvidia to the prevention and control of water inrush in coal mines.This paper introduces the idea and model of CUDA water inrush prediction software in detail.The CUDA monitoring water inrush program is compiled and the actual prediction of water inrush points in the coal mine is carried out in combination with the drilling water level data.Compared with the traditional CPU serial calculation,the calculation speed of water inrush prediction in the coal mine is greatly improved,which has reference significance for realizing fast and accurate prediction of water inrush in coal mines.