Aiming at solving the problem of surfactant-enhanced DNAPLs-contaminated aquifer remediation, based upon multiphase flow numerical simulation model in this study, the Latin hypercube sampling(LHS) method used for collecting input-output sample data of the multiphase flow simulation model, to effectively improve the efficiency of random sampling and coverage. According to the inputoutput sample data sets, multiple regression analysis method was used to build an alternative model,dual response surface model, for the multiphase flow simulation. The study explores new theories and methods for optimization of DNAPLs-contaminated aquifer remediation. This study selected a new set of schemes to test the accuracy of the alternative model. The relative errors of the results calculated for the alternative model are all less than 10％, which means the DNAPLs-contaminated aquifer simulation model based on the LHS method is much close the simulation model. The alternative model can significantly reduce the huge computational load caused by using simulation model directly and repeatedly during the optimization process.%针对表面活性剂强化的重非水相流体(DNAPLs)污染的含水层修复问题,在建立多相流数值模拟模型的基础上,应用拉丁超立方采样(LHS)方法,在多相流模拟模型可控输入变量的可行域内采样,有效提高了采样效率和覆盖程度.根据采集的样品数据集,运用多元回归分析方法建立多相流模拟模型的替代模型--双响应面模型,为DNAPLs污染含水层修复过程的优化设计的耦合技术探索新的理论和方法.经栓验,替代模型计算结果的相对误差均小于10%,精度较高,说明其在功能上充分逼近模拟模型.运用替代模型实现模拟模型与优化模型的连接,可以大幅度减少优化模型计算过程中直接多次反复调用模拟模型所引起的庞大计算负荷.
展开▼
