Determining the levels and parameters of thief zone based on automatic history matching and fuzzy method

Abstract

Abstract Thief zone which evolves from long-term water flooding, has become a subject of concern for reservoir engineers, as they lead to early water breakthrough in oil producers and uneven sweep around water injectors, thus it is essential to select wells which need to be modified injection or production profiles. This article presents a methodology of determining the levels and parameters of thief zone in different areas of the reservoir on the concept of “based on the information from every grid of reservoir model” by using automatic history matching and fuzzy method. Since characterizing the reservoir uncertainty is crucial to the reservoir description and future performance predictions, automatic history matching using ensemble Kalman filter (EnKF) with covariance localization is first proposed. Then according to theory of logical analysis (TLA), fuzzy analytic hierarchy process (FAHP) and Fuzzy comprehensive evaluation (FCE) method, the system to quantitative evaluation of thief zone is presented, and the reservoir can be graded into three categories of severe thief zone, light thief zone and no thief zone. The methodology has been applied to X oilfield in western North China which has 17 layers from the top to the bottom in the stratigraphy, and the results show that 5 layers exist severe thief zones and the volume of severe thief zones is the largest in layer 32, and there are four wells in this layer that their injection or production profiles must be modified. In addition, the interwell tracer test result shows the proposed methodology is more accurate by comparing with other methods in the references which mainly rely on the properties of single well to determine the levels of thief zone. The proposed approach is more accurate and less manpower needs to identify thief zones, which also providing a strong basis for oilfield development adjustment in high water cut stage.