摘要
在分析影响回采工作面来压显现主要因素的基础上,通过LevenbergMarquardt优化算法来改进人工神经网络,以现场实测工作面支柱支护阻力、超前支护下沉量和推进步距等作为样本值,建立了综采工作面矿山压力来压预警预报神经网络模型,并以此为核心设计了回采工作面预报预警系统,该系统包括工作面及两巷数据录入与管理、数据综合处理、数据动态曲线分析、压力预报预警、数据报表打印等功能。并利用该系统对枣庄矿业集团新安煤矿新源井12101回采工作面来压进行预报预警,正确地预报了工作面的来压及各种安全因素的预警,避免了回采工作面安全事故的发生。
On the basis of analyzing the factors which influence the ground pressure of mining working, an improved BP neural network is set up which adopts Levenberg Marquardt optimized algorithm. In the algorithm, the sample values are composed of support capacity of support resistance of working surface, sinking measures of pre-support and pressing distance of working surface. We set up a forecasting and pre - alarming system of working surface pressure hased on this model. In this system, the following functions are included ： data input and management, synthetic data processing, dynamic curve analysis of data, forecasting and pre - alarming for workface and road pressure and printing data report. It is used to forecast the ground pressure on 12101 mining working surface of Xinan Mine of Zaozhuang Mine Group, and the forecasting result proves that this method is very useful.
出处
《计算机仿真》
CSCD
2006年第11期287-290,共4页
Computer Simulation
关键词
神经网络
算法
矿山压力
预警预报
Neural network
Algorithm
Ground pressure
Forecasting and pre - alarming
