基于遗传算法的非饱和溶质运移参数优化

根据室内实验和数值模拟数据,分别以土壤含水率和溶质质量浓度的实测值与其计算值标准差最小为优化目标,建立溶质和水分运移参数识别的多目标优化模型,采用遗传算法、有限差分法和线性加权法相结合的计算方法,对模型进行优化求解。分别以单组、2组和3组时刻实验数据组合作为参数优化初始数据,获取了土壤非饱和水分和溶质运移参数值,并以1 345 min时刻实验数据进行验证。含水率计算值与实测值相关系数的最大值和最小值分别为0.975 3和0.945 0;溴离子溶液质量浓度的计算值与实测值相关系数最大值和最小值分别为0.964 6和0.935 2,实测值与计算值吻合较好,表明用这一方法识别非饱和水分-溶质运移参数是可行的。

Optimization of Solute Transport Parameters in Unsaturated Soil Based on Genetic Algorithm

According to the experiment and numerical simulation, a multi-objective optimization model for identification of solute and water transport parameters was built up, which use soil water content and the minimum standard deviation of measured and observed water content and solute as optimized objective. A combination of genetic algorithm, finite difference and linear weighted method was applied for calculation. Taking one, two and three groups experimental data as initial data, solute transport parameters in unsaturated soil were obtained and verified by using the experimental data at 1345min. The maximum and minimum correlation coefficient between calculated and measured value of water content were 0.9753 and 0.9450. The maximum and minimum correlation coefficient between calculated and measured value of bromonium ion concentration was 0.9646 and 0.9352. The measured results were well coincided with the calculated results and showed the feasibility of the proposed method.