摘要
构建混合改进的遗传算法(GA),实现F-X模型参数反演,以期为非饱和渗流机制研究提供指导。以F-X模型中的系数a、m、n以及饱和体积含水量θ_(s)作为模型参数,以简单十进制遗传算法为基础,融入优体克隆操作等算子,形成混合改进的遗传算法;采用已有数据进行算法验证,对不同干密度土样的F-X模型参数进行反演,明确模型参数与干密度间的相关关系。研究发现:1)混合改进的遗传算法反演F-X模型参数精度较高,模型拟合结果与实测值均方差均小于0.03;2)随着土样干密度的增大,F-X模型的参数a是控制土—水特征曲线(SWRC)的主要因素,而参数m、n变化不显著;3)建立的F-X模型参数与干密度间的关系也可广泛适用于其他土样的SWRC,大大简化了模型参数的获取过程。
A hybrid improved genetic algorithm(GA)was constructed to realize the inversion of F-X model parameters in order to provide guidance for the study of unsaturated seepage mechanism.Taking coefficients a,m and n in the F-X model and saturated volumetric water content θ_(s)as model parameters,based on simple decimal genetic algorithm,and incorporating operators such as optimal body cloning operation,a hybrid improved genetic algorithm was formed.The existing data were used to verify the algorithm,the F-X model parameters of soil samples with different dry densities were inverted,and the correlation between model parameters and dry density was made clear.The results are as follows:1)The hybrid improved genetic algorithm has high precision in the inversion of the F-X model parameters,and the mean square deviation of model fitting results and measured values is less than 0.03.2)With the increase of dry density of soil samples,the parameter a of F-X model is the main factor to control the Soil-Water Retention Curve(SWRC),while the parameters m and n do not change significantly.3)The relationship between F-X model parameters and dry density can also be widely applied to SWRC of other soil samples,which greatly simplifies the process of obtaining model parameters.
作者
李昱宛
LI Yuwan(CSSC Institute of the Engineering Investigation&Design Co.,Ltd.,Shanghai 200333,China)
出处
《工业技术创新》
2021年第6期113-118,共6页
Industrial Technology Innovation
