基于漂移度的隧道涌水量组合解析预测方法

隧道涌水量解析预测方法众多,各种解析法预测结果往往存在不一致性。为解决上述问题,引入组合预测思想,采用线性加权组合形式对常用的隧道涌水量解析计算方法进行组合。基于"少数服从多数"的原则,以目标方法外其他方法的整体预测趋势作为参考系,引入漂移度指标对各单项解析法进行可靠性评价,运用合理的权重分配方式建立了组合解析预测方法。以广州某地铁隧道工程为研究对象,选用了裘布依模型等4种常用的解析法对隧道正常涌水量分别进行了单项和组合预测。结果表明:在最佳预测方法未知的情况下,组合预测方法能较为准确地预测隧道实际涌水量,验证了该方法的可靠性。

Combined forecasting method of tunnel gushing water based on drift degree

A number of analytical methods have been proposed to forecast tunnel gushing water so far. However, the forecast results are always inconsistent. In order to solve these problems, several analytical methods were combined by linear weighting according to the combination forecasting consideration. The drift degree was introduced to estimate the reliability for each analytical method with the reference of the general tendency of other methods based on a principle of minority being subordinate to the majority. Then a rational weight distribution method was applied to develop a combination model composed of multiple analytical methods. Taking a metro tunnel in Guangzhou as the research object, the single and combined predictions for tunnel gushing water were made based on four kinds of common analytical methods such as Jules Dupuit model. The results show that the rushing water calculated by proposed method matches the actual data, implying that the proposed method is reliable.