基于线性二次指数平滑法的瓦斯含量预测

Gas Content Prediction Based on Linear Double Exponential Smooth Method

为了提高煤层瓦斯含量的预测精度,以某矿区垂深间隔为50 m的瓦斯含量数据,瓦斯含量运用线性二次指数平滑法建立瓦斯含量预测模型,并结合折线图和方差分析比较预测结果与实测数据。结果表明:平滑常数α为0.8的线性二次指数平滑法模型预测精度较高,运用该模型既能满足瓦斯含量随埋深增加呈线性规律的总体趋势,又能根据最新实测瓦斯含量比较准确地预测煤层瓦斯含量;依据实测值参与模型的权数规律,模型能够接近最新一期的实测数据。

In order to improve the accuracy to predict gas content of seam,based on gas content in vertical depth with interval of 50 m in a mining area,a linear double exponential smooth method was applied to establish prediction model to predict gas content and in combination with polygonal line graph and variance analysis,a comparison was conducted on the predicted results and the measured data. The results showed that the linear double exponential smooth method model with a smooth constant α of 0. 8 could have high prediction accuracy. The application of the linear double exponential smooth method to predict the gas content could meet a general tendency of the gas content increased in linear law with the depth increased and also could accurately predict the seam gas content based on the comparison on the updated measured gas content. Based on the measured values involved in the weight law of the model,the model could approach the measured data of a new round.