基于TLBO-LOIRE的回采工作面瓦斯涌出量预测

Gas Emission Quantity Prediction of Working Face Based on TLBO-LOIRE Method

瓦斯涌出量是瓦斯防治与管理、矿井通风系统设计的重要基础数据,准确地预测瓦斯涌出量对于煤矿安全生产有着极其重要的指导意义与应用价值.但工作面瓦斯涌出规律复杂,在检测、数据采集过程中不可避免地会混入异常噪声,直接影响着瓦斯预测的准确性.本文采用l1正则化异常值隔离与回归方法(LOIRE)对煤矿回采工作面瓦斯涌出量及其相关影响因素的统计样本数据库进行计算分析,隔离样本的异常噪声干扰,利用教与学算法(TLBO)优化回归参数,建立了回采工作面瓦斯涌出量的优化预测模型,并对煤矿现场数据进行分析预测,结果表明3个回采工作面的瓦斯涌出量预测误差分别为3.04%、0.33%和2.36%,平均相对误差仅为2.36%.TLBO-LOIRE优化预测方法,预测准确性高,能够满足井下瓦斯防治的工程需要,对其它工程领域的数据预测同样适用.

Gas emission quantity is an management and ventilation system important basic data of gas design in coal mine, therefore emission prevention and accurate prediction of gas emission quantity is of great significance and application value. But the gas emission law is very complicated, and the gas data mixed with some abnormal noises inevitably in detection and data acquisition would affect the prediction accuracy of gas emission directly. In this paper, l^- regularized Outlier Isolation and Regression （ LOIRE ）, which can isolate the abnormal noise interference of samples, is applied in calculating and analyzing the statistical sample database of gas emission quantity and its related influence factors in working face of coal mine. And the penalty coefficient A of LOIRE method is optimized through the Teaching Learning Based Optimization （TLBO） algorithm. Moreover the optimal prediction model of gas emission quantity is established by TLBO-LOIRE. Then the sample data of gas emission is predicted and analyzed, the calculation results show that the relative errors of gas emission quantity prediction of 3 working face are 3.04% ,0. 33% and 2. 36%, respectively, and the mean relative error is only 2.36%. The TLBO-LOIRE optimal prediction method with high accuracy meets the needs of gas prevention and control engineering applications in coal mine, also can be applied to some other engineering fields for data prediction.