【摘要】 为有效分析煤矿瓦斯监测数据,较准确地预测瓦斯浓度,根据矿井瓦斯实时监测数据和现场采煤方式的特点,分解数据形成的时间序列,用自回归(AR)、径向基函数神经网络(RBFNN)和高斯过程回归(GPR)3个模型组合进行预测。以预测有效度作为其精度的评估指标,得出最佳瓦斯浓度预测结果,实现瓦斯浓度自适应预测,并结合瓦斯监测数据的统计特征,实现瓦斯浓度的实时动态预警。实例分析表明:应用该方法能够提高预测精度,实现超前预警。更多还原

【Abstract】 For the purpose of achieving more accurate prediction of gas(methane) concentration in coal mine through effective analysis of the gas monitoring data,the characteristics of real-time monitoring data and the mining way in coal mine production were analyzed.Thus gas monitoring data time series was decomposed.AR,RBFNN and GPR were integrated for prediction.The effectiveness of gas concentration prediction was taken as the criterion for assessing accuracy of prediction.An adaptive gas concentration prediction was achieved.Dynamic gas pre-warning analysis was achieved.Results of a case study on a certain coal mine show that with this method,the prediction accuracy can be improved and a pre-warning can be issued.更多还原