The use of residual polymer in the formation after polymer flooding for deep injection-profile control is an important method to further improve oil recovery. In order to evaluate the sensitivity of profile control by clay particles after polymer flooding, 5 influential factors were selected to perform the numerical simulation, including the permeability ratio, formation oil viscosity, injection timing, injection volume and injection concentration of clay particles solution. The results show that, with the increase of such factors as permeability ratio and formation oil viscosity, the recovery increment of the profile control increases first and then decreases; the earlier the clay particle solution is injected, the better the development effect; the larger the injection volume and injection concentration of clay particles solution, the larger the recovery increment of the profile control. After that, Response Surface Methodology was used to conduct the performance prediction and parameters optimization. Firstly, the Box-Behnken model with three levels was adopted to perform the experiment design. Then, two predictive models of ultimate recovery and recovery increment of the profile control were established on the basis of the multiple regression analysis, which agrees well with the numerical simulation results. Finally, the optimal level of each influential factor was determined by response surface method.

聚合物驱后在地层中使用残余聚合物进行深层注入剖面控制是进一步提高采油率的重要方法。为了评价聚合物驱后粘土颗粒控制剖面的敏感性，选择了5个影响因素进行了数值模拟，包括渗透率，地层油粘度，注入时机，注入量和注入量。结果表明，随着渗透率，地层油黏度等因素的增加，调剖剂的采收率先增加后降低。越早注入粘土颗粒溶液，显影效果越好；粘土颗粒溶液的注入量和注入浓度越大，概要文件控件的恢复增量越大。之后，使用响应面方法进行性能预测和参数优化。首先，采用三个层次的Box-Behnken模型进行实验设计。然后，在多元回归分析的基础上，建立了轮廓控制的最终采收率和采收率恢复增量的两种预测模型，与数值模拟结果吻合较好。最后，通过响应面法确定各影响因素的最佳水平。在多元回归分析的基础上，建立了轮廓控制的最终采收率和采收率恢复增量的两个预测模型，与数值模拟结果吻合良好。最后，通过响应面法确定各影响因素的最佳水平。在多元回归分析的基础上，建立了轮廓控制的最终采收率和采收率恢复增量的两个预测模型，与数值模拟结果吻合良好。最后，通过响应面法确定各影响因素的最佳水平。