基于感知生成对抗网络的工程地质剖面图生成方法研究

fiResearch on Engineering Geological Profile Generation Method Based on Perceptual Generative Adversarial Networks

工程地质剖面图是地下工程不可或缺的重要组成部分,也是施工和运营期间变形监测与安全预警的基础之一。其传统绘制方法不一定能完全反映出地层空间分布的随机性,这对工程实际地质剖面的认知造成了极大不利影响。现有工程地质剖面图的绘制一般由工程师将钻孔和实验数据导入对应软件中,借助软件的数学插值方法进行绘图,这一过程很难体现出土壤性质的空间变异性。本研究提出了一种基于感知生成对抗网络的新型工程剖面绘图方法:以真实地铁监测数据为例,构建了输入输出成对数据,通过生成模型自动生成工程地质剖面图。该方法在指定条件下与其他多种相关算法进行对比,均取得了较优的效果,具有一定的实际应用价值。

Engineering geological profile is an indispensable part of underground engineering,and is also one of the bases of deformation monitoring and safety warning during construction and operation.The traditional drawing method may not fully reflect the randomness of spatial distribution of strata,which has a great negative impact on the perception of the actual engineering geological profile.Typically,engineers import borehole and experimental data into corresponding software and then utilize mathematical interpolation methods in the software for drawing.However,it is difficult to reflect the spatial variability of soil properties in this process.In this study,a new engineering profile mapping method based on perceptual generative adversarial networks is proposed:taking real subway monitoring data as an example,input-output paired data are constructed,and the engineering geological profile is automatically generated by generative modeling.The method is compared with a variety of other related algorithms under the specified conditions,all of which achieve superior results and have certain practical application value.