论文部分内容阅读
鉴于使用确定性模型预测隧道涌水量时存在难以准确获取水文地质参数等诸多不便,本文将地下水系统视为“黑箱”模型,通过提取隧道涌水量历史观测数据本身蕴涵的趋势、周期和随机变化规律,建立了隧道涌水的时间序列预测模型。经用于铜锣山隧道实例,反演系列的平均绝对误差为14.67%,预测序列的平均绝对误差为14.34%,表明模型的整体预测精度较高。
In view of the inconvenience that it is difficult to accurately obtain the hydrogeological parameters when using the deterministic model to predict the water inflow in the tunnel, this paper considers the groundwater system as a “black box” model. By extracting the trend, period and random The time series forecasting model of gushing water of tunnel is established. After being used in the tunnel of Tongluo Mountain, the average absolute error of the inversion series is 14.67% and the average absolute error of the prediction sequence is 14.34%, indicating that the overall prediction accuracy of the model is high.