Research On Integrated Kalman Filtering Method For Data Integration In Oilfield Development

With the development of oilfields enter in the mid and later period, a big problem which need to be addressed urgently is to enhance oil recovery efficiency. Production optimization is one of the important way to increase production of oil and gas. However, for production automatical opitimation technical, the key point is how to integrate data from different source into the production optimize dynamic updating modle, which cope with the reservoir parameters and geologeical model. Therefore, choose right integration technology and method is particularly important in reservoir dynamic development.This article utilize ensemble Kalman filter (Ensemble Kalman Filter, referred to EnKF) to create a new automatical matching reservoir model,and consider each of the reservoir parameter as a collection vector, match real-time monitored data(such as’rate、pressure、liquid withdrawal plane’)and constantly updates reservoir model state parameters (porosity、 pemeability、water saturation)to approach the true value of the reservoir gradually,finally make the prediction consist with production dynatic. Set up the target function based on the difference between monitored data and caculated data for ensemble Kalman filter algorithm optimization method, which aid reservoir history matching; use Monte Carlo method to generate a set of initial collection of reservoir model,combining the ensemble Kalman filter method to match the monitoring data, and get a better reservoir model. This paper also did the real-time update prediction of reservoir parameters (including static parameters such as porosity; dynamic parameters such as water saturation), discussed the influence of collection size, noise ratio, the mean initial value of sample on the reservoir history matching results; compared the difference between the production data and monitored data; carried out the prediction uncertainty analysis on reservoir model parameters(such as cumulative oil, gas/oil ratio,ect). Based on what we have done, take the typical five-spot network as example,apply the automatic matching reservoir model of ensemble Kalman filter to real-time update the reservoir parameter, which verified the method presented by this aritical.The viewpoint presented in this paper have a instructional meaning on reservoir exploitation and data integration during reservoir development.