摘要
为准确预测矿井涌水量,针对矿井涌水的动态混沌特征,将混沌方法应用到矿井涌水量预测中,其预测步骤为:收集矿山日涌水量时间序列,运用C-C方法确定其重构参数,对其进行相空间重构。以定性和定量方法结合辨识重构序列的混沌特征。在此基础上,根据加权一阶局域法建立矿井涌水量预测模型。以某矿山-40 m中段日涌水量时间序列为例,计算其理论预测时长为10天,按混沌方法对序列最后20组数据进行预测检验。对于前10组数据(理论预测时长内),其预测平均误差为3.63%,最大预测误差为7.42%;对后10组数据,其预测平均误差为5.14%,最大预测误差为14.11%。研究结果表明,将混沌方法应用于矿井涌水量预测中是可行的,但仅能实现短期预测。
In view of the limitations existing in the prediction of mine water inflow by traditional meth- ods, a chaotic method was worked out and used to predict mine water inflow. The prediction procedure was： daily mine water inflow time sequences were collected, their reconstruction parameters were deter- mined by C - C method, and the phase space was reconstructed. The chaotic characteristics of reconstruc- ted sequence were identified by the combination of qualitative and quantitative methods. A mine water inflow prediction model was built by adding_weight one_rank local_region method. Taking a daily water inflow sequence of a - 40 m mine area as example, its theoretical forecast length was calculated as 10 days, the last 20 data on the water inflow sequence were used for prediction test. The average forecast error of the first 10 test samples （theoretical forecast length） was 3.63% , the maximum forecast error was 7.42%; the average forecast error of last 10 samples was 5. 14%, the maximum forecast error was 14.11%. The research results show that applying chaotic method into mine water inflow forecast is feasi- ble, but only the short-term forecast could be realized.
出处
《中国安全科学学报》
CAS
CSCD
北大核心
2013年第4期51-56,共6页
China Safety Science Journal
基金
国家自然科学基金资助(51274250)
关键词
矿井涌水量
混沌
C-C方法
相空间重构
加权一阶局域法
mine water inflow
chaos
C - C method
phase space reconstruction
adding_weight one_rank local_region method
