基于可靠度理论的地震滑坡运动距离预测模型

滑坡运动距离是评价滑坡致灾范围的一个重要指标,因此其预测方法一直备受关注。实际滑坡运动距离往往具有很强的随机性,而目前的预测模型往往给出一个确定性的计算公式,未考虑其随机性。基于此,首先,通过对前人提出的基于能量守恒原理的滑坡运动距离理论模型进行分析,认为滑坡运动距离与滑坡体积V和滑坡前后缘高差H成正相关,由此提出滑坡运动距离的一般函数关系式; 其次,基于“5·12”汶川地震诱发的46个滑坡案例,对影响滑坡运动距离的主要因素进行了相关性分析,认为滑坡体积V、滑坡前后缘高差H及其组合VH对滑坡运动距离影响显著,并由此建立了滑坡运动距离的多元回归统计模型; 最后,根据工程重要性的不同,在上述研究的基础上提出了基于可靠度理论的滑坡运动距离统计模型。算例分析表明,置信度越高,滑坡运动距离预测范围就越大,相应的搬迁和防护费用就越高,因此应根据实际情况,选择合理的置信度。研究成果有助于进一步优化滑坡运动距离预测的统计模型,进而为滑坡灾害评估和防治提供可靠依据。

Prediction model for the landslide movement distance induced by earthquake based on the reliability theory

The landslide movement distance is an important index to evaluate the scope of landslide hazard, therefore its prediction method is always being concerned. The actual landslide movement distance often has a strong randomness, but a deterministic calculation formula is offered in current prediction models without the consideration of the randomness. Based on this situation, Firstly, the theoretical model for the landslide movement distance proposed by predecessors was analyzed, and the landslide movement distance is mainly positively correlated with the landslide volume V and elevation H. Then, the authors proposed the general functional relation for the landslide movement distance. Secondly, based on 46 landslide cases induced by “5·12” Wenchuan Earthquake, the correlation analysis of the main factors affecting the landslide movement distance was carried out. It is concluded that the landslide volume V, elevation H and their combination VH have significant influence on the landslide movement distance, and the multiple regression statistical model for the landslide movement distance was established. Finally, the statistical model for the landslide movement distance based on the reliability theory was put forward, according to different engineering importance. The calculation example shows that the higher the confidence is, the larger the predicted range of the landslide movement distance will be, with the higher the relocation and protection cost. Therefore, a reasonable degree of confidence should be chosen according to the actual situation. This study is helpful to further optimize the statistical model for the prediction of the landslide movement distance, and then provides reliable evidence for landslide disaster evaluation and control.