降雨影响下滑坡变形预测的GA-NN模型研究

Study on GA-NN model for forecasting the displacement of landslides affected by rainfall

八尺门滑坡是福-宁高速公路施工现场的一个降雨型滑坡,近年来该滑坡由于人为及降雨的因素而频频发生滑动,对坡下的高速公路施工及滑坡区内的人民生命财产安全造成了严重威胁。为了准确掌握降雨量与滑坡位移之间的关系,在对八尺门滑坡现场降雨量及滑坡位移的长达一年的监测资料的基础上,利用GA-NN模型,对降雨条件下的滑坡位移演化发展做出了精确的预测。研究成果已作为重要的基础性数据被有关施工和研究部门采用,对于指导现场施工和保证滑坡影响区内人民生命财产的安全发挥了重要作用。

Bachimen landslide lies in a highway construction site, and hundreds of people are living nearby. In recent years, by human factor and the rainfall, the slide became moving frequently and threatened the residents and workers on it. Rainfall is the key factor leading to the landslide, so study on the relation between rainfall and displacement of landslide is essential to forecast the landslide. Combining genetic algorithm and neural network, a GA-NN model is proposed whose structure is optimized. The relation between the rainfall and the displacement of landslide is mapped by this GA-NN model. The research result has been adopted by the highway construction corporation and certain institutes, and it was very useful for directing the highway construction and ensuring the safety of the residents?lives and wealth.