摘要
针对现阶段瓦斯涌出量预测中存在的样本数据库过饱和现象,提出了一种限定记忆模式的多维GM-RBF瓦斯涌出量预测模型;基于软测量思想引入了代谢因子,变一维瓦斯涌出量数据为多维"辅助变量"和"主导变量",构建了多维动态数据集;对车集煤矿2612工作面的实例验证结果表明:限定记忆模式下的多维GM-RBF模型拟合曲线离散性最小,瓦斯浓度变化趋势和实际监测结果最为接近,对煤矿工作面瓦斯涌出量的预测具有更高的准确性。
In allusion to the problem of data supersaturation during gas emission prediction at the present stage, an on-line gas emission prediction model based on GM-RBF of limited memory is put forward. Based on soft measuring thought, metabolic factor is introduced to change one-dimensional index variable of gas emission into multi-dimensional auxiliary variable and dominant variable. A case study in No. 2612 working face of Juji coal mine is implemented. Results show that the proposed model has the minimum discreteness. The prediction results are closest to the actual values, indicating that the proposed model is able of predicting gas emission constantly and accurately.
作者
李俊哲
秦志
周鑫隆
LI Jun-Zhe;QIN Zhi;ZHOU Xin-long(School of Resources and Environmental Engineering,Wuhan University of Technology,Wuhan 430070,China;Juji Coal Mine,Yongcheng Coal and Electricity Holding Group Co.,Ltd.,Yongcheng 476600,China)
出处
《煤炭技术》
CAS
2019年第1期92-95,共4页
Coal Technology
