聚合物驱含水动态规律的ARMAX法建模与预测

用数值模拟方法研究聚合物驱油动态规律由于影响因素多且不确定，得到的结果与实际效果相比会有一定偏差，而利用历史数据建模时大多只考虑综合含水、注入黏度、渗透率等参数，而没有考虑反映聚合物驱的开发生产和注入参数，且建立的多为静态模型，不能完全反映聚合物驱的动态规律。为此，提出了一种系统辨识理论的建模方法，对油田聚合物驱的含水变化进行预测。该方法建立的模型是一带有外部激励的自回归滑动平均模型（ARMAX）。所建立的模型主要考虑了注入浓度、采油速度等影响聚合物驱含水变化的动态因素。通过实际聚合物驱动态数据测试，验证该方法建立的模型预测精度较高。利用ARMAX模型对聚合物驱含水变化进行预测是评价聚合物驱效果并降低风险的一种有效的方法。

Modeling and prediction of water cut performance for polymer flooding based on ARMAX

A new modeling method is presented based on system identification theory to predict the water-cut variation of the polymer flooding. Dynamic factors of the polymer flooding exploitation are considered in the modeling procedure, and the model is an autoregressive (AR) moving average (MA) model with exogenous (X). This paper develops an ARMAX model for predicting water-cut change curve of polymer flooding which usually looks like a funnel, and the dynamic history data are used directly to establish regression model. The model considers many dynamic factors affecting water-cut of polymer flooding such as injecting concentration, oil production rate etc. Practical dynamic data verified that the prediction results had high precision. The ARMAX prediction model is a powerful tool for evaluating the effects and reducing the risk of polymer flooding.