Polymerflooding Dynamic Index Forecasting And Its Estimation

As the largest oilfield in China, the production of polymerflooding in Daqing Oilfield has exceeded more than 10,000,000 tons for over 9 years continuously, which is more than 25% of total production per year of the whole Daqing Oilfield. Moreover, Daqing Oilfield has been in the extra-high watercut stage, to achieve the strategic target of 40,000,000 tons per year continuable production in Daqing Oilfield, polymerflooding production is going to be the recent main measurement of increasing production.Forecasting is the base and precondition of decision. The development impact estimation, development programming, development plan design and adjustment of an oilfield are all based on the dynamic index forecasting (DIF). Therefore, the DIF of polymerflooding is the focus of oilfield.For polymerflooding dynamic forecasting and its estimation, this paper researches as follows:(1) A new clustering method-IFKN net has been founded for the clustering calculation of polymerflooding blocks. Because the single data in FCM has attracted all nodes, the attractions can be adjusted even the cluster center is far, and there is no dead nodes, which improves the optimization probability and reduces the probability of slumping local extremum. Because of the continuity of geologic sediment process in Daqing Oilfield, the blocks with polymerflooding are not only different but also connected closely, and the FCM method will be slow for this kind of clustering calculation. FCM method is improved by introducing Mexico straw function, and expert experiences has been taken into forecasting process, which not only improves the speed, but also guarantees the optimization probability.(2) The research results of oil displacement mechanism of polymer solution from elder experts are summarized, especially for the viscoelasticity of polymer. Under the precondition of ensuring the mechanism of polymer solution, the changing rules of dynamic index in the actual block with polymerflooding are analyzed. By making the simple experience plate of polymerflooding dynamic index, the primary forecasting work of polymerflooding index can be done easily.(3) Analyze the single factor of polymerflooding dynamic forecasting. By using reservoir simulation software, multi-concept geological models are founded, and considering geological factors (including heterogeneity, connection style and thickness proportion of channel sand), development factors (including polymer dosage, injection speed, injection concentration, primary watercut before injecting polymer, corresponding index of well team, well space) and the characters of polymer solution (viscosity, shear thinning, adsorption, dead pore volume and residual resistance factor) for the effects to the dynamic index of polymerflooding, therefore, which is certainly be helpful for the work of dynamic index forecasting.(4) "Dynamic combined forecasting method of polymerflooding" is founded. Under the guide of combining forecast thought, the front movement equations of polymerflooding and regression are combined together, considering the character of dynamic index that varying rapidly of polymerflooding, thereby, not only can experience regression guarantee higher forecasting precision for fluid production, but also response for instantaneous change in time by using the front movement equation, and finally, the average precision is improved by 6%.(5) The former integrated dynamic analysis method considered geological factors, development factors and production factors during the polymerflooding, but the watercut curve and liquid curve were linearity regression, this would produce about 5% error. "Counter-supply oil production" was put forward in this paper, sine function was used to describe the oil production during the low watercut stage, and the result would be recalculated to the oil production by intergrated dynamic analysis method, this method improves the precision by 5% than that of integrated dynamic analysis method.The two methods are used in polymerflooding blocks, the results shows that, the precision is improved greatly, and could meet the needs of field work, moreover, applying these methods in the continuously production program of 40,000,000 tons in Daqing Oilfield solves the status of hard programming of polymerflooding index.