Combination of Genetic Algorithm and Inverse Solution Technique for Estimating the Hydraulic Properties of Unsaturated Soft Rock

Abstract

Unsaturated hydraulic property (relation between the water retention curve and hydraulic conductivity function) of soft rock must be properly estimated for analyzing the saturated-unsaturated groundwater flow such as the flow around a tunnel, rock surface and so forth. In this study a practical technique for estimating the unsaturated hydraulic property of soft rock was developed on the basis of the Campbell model and Genetic Algorithm (GA). The major objectives of the present study are the following two; 1) investigation whether or not the Campbell model can be applied to soft rock for approximating the unsaturated hydraulic properties and 2) application of GA technique for estimating two parameters included in the Campbell model. These parameters were estimated through the inverse analysis of the transient change of evaporation rates from a surface of disk shaped specimen of Tertiary sedimentary rock which are sampled from the Rokkasho area, Aomori prefecture. Two transient evaporation experiments under different atmospheric conditions were proposed for the estimation. The Campbell's parameters of 13 soft rock specimens (5 sandstone and 8 tuff breccia specimens) were estimated through those types of experiments. As the results it was clearly found that the transient change of evaporation rates can be well fitted by the Campbell model except for the region of low saturation and this fact implies that the Campbell model can be applied to the soft rock. And also it was found that, the GA is a promising technique for parameters estimation in the inverse solution method. It can be concluded that the techniques are practically for estimating the unsaturated hydraulic properties of soft rock.