摘要
边坡失稳风险评估是防范滑坡灾害发生的有效方式之一。为提升岩质边坡失稳风险评估的客观性、有效性和可靠性,基于深度学习栈式自编码网络,提出一种考虑岩体质量的岩质边坡失稳风险评估方法。该方法使用栈式自编码器网络自动计算出边坡岩体质量等级,再将岩体质量等级、坡高、坡角等8个指标作为输入参数,基于样本数据集构建岩质边坡失稳风险等级评估的栈式自编码器网络,并在基于误差反向传播算法的多代训练中形成最优、高精的风险评估网络。案例应用表明:提出的考虑岩体质量的岩质边坡失稳风险评估方法可准确地评估岩质边坡失稳风险等级,预测准确率达85%以上;该方法考虑了岩体质量等级、地形地貌、降雨等客观因素对岩质边坡稳定性的影响,实现了客观、科学、快捷的岩质边坡失稳风险评估。
Instability risk assessment of slope is an effective way to prevent landslide.In order to improve the objectivity,validity and reliability of rock slope instability risk assessment,the research of instability risk assessment of rock slope was carried out based on the stacked autoencoder network.Firstly,the stacked autoencoder was used to calculate the rock mass quality grade of the slope,and sample data set including 8 factors such as the rock mass quality grade,slope height,slope angle,and so on were used to construct the stacked autoencoder network for instability risk assessment of rock slope.In the multi generation training based on error back propagation algorithm,an optimal and high-precision risk evaluation network was formed.The results show that the proposed risk assessment method for rock slope can accurately assess the risk level,with an accuracy rate of more than 85%.The proposed method fully considers the impact of rock mass quality,terrain,landform,rainfall,and other influencing factors on the slope stability.It realizes an objective,scientific,and fast risk assessment of rock slope.
作者
于锦
熊齐欢
谭飞
樊孝富
曹军龙
YU Jin;XIONG Qihuan;TAN Fei;FAN Xiaofu;CAO Junong(CCCC Second Harbor Engineering Co.,Ltd.,Wuhan 430048,China;Faculty of Engineering,China University of Geoscience(Wuhan),Wuhan 430074,China)
出处
《安全与环境工程》
CAS
CSCD
北大核心
2023年第4期110-119,共10页
Safety and Environmental Engineering
基金
国家自然科学基金项目(41672260)
云南省科技创新示范项目(云交科教[2019]15号)。
关键词
岩质边坡
失稳风险评估
栈式自编码器
岩体质量
rock slope
instability risk assessment
stacked autoencoder
rock mass quality
