Identification Of Dominant Seepage Channels In Waterflooding Development Reservoirs

The Bozhong 34-1 oilfield is located in the central structural ridge of the Yellow River Sag in the southern Bohai Sea.The main production interval,with high porosity and high permeability,is sandstone of the shallow water delta and the meandering river sedimentary in the lower section of the Minghuazhen Formation.The oilfields are produced by water injection development and have reached the stage of high water cut development.The existence of reservoir heterogeneity causes the fluid flow differently and causes most of the fluid to flow along the channels where the seepage resistance is relatively small.The channel with relatively small seepage resistance is the dominant water flow channel.The existence of the channel causes serious flooding of the reservoir,and causes a lower utilization rate of the injected water,lower sweep coefficient,decreased production of the single well,a rapid increase of water content,and so on,which restrict the improvement of the production efficiency of the oil field.In order to achieve the objective of stable production increase in the Bozhong 34-1 oil field,it is necessary to identify and predict the dominant seepage channel of the oilfield waterflood,in order to provide the basis for the subsequent production adjustment and reform the unformed interval.Based on the theory of comprehensive judgment,this article takes the mathematical geology and development geology as the guide.The article selects the research method to identify and predict the dominant percolation channel by analysing the geological factors and production factors of the dominant flow channel.Firstly,using dynamic data and production test data to ensure the single well which exists the dominant water-flooding channel by the method of fuzzy comprehensive.Then,on this basis,statistically analyze the characteristics of logging curves and geologic parameters for each production interval of a single well of a waterflood dominant percolation channel,and comprehensively evaluate the reservoir characteristics of the production horizon to determine the most likely formation.At the same time,the relationship between sand bodies in injection and production wells is combined to complete the identification of the direction of the dominant percolation channel for waterflooding.In order to be able to quickly and accurately complete the identification and prediction work,this article is guided by deep learning theory.The recognition results of comprehensive recognition are used as learning training samples.Deep learning is performed using BP artificial neural network to obtain the influencing factors and the formation of water drive advantage percolation channels.Through the analysis and research,the following understandings have been obtained:1)The water injection dominant flow channel is formed by water injection well with a fuzzy comprehensive identification index greater than 0.6 and the production well with a fuzzy identification index greater than 0.7;2)The effect degree of dynamic parameters on the formation of dominant water flow channels is higher than the static parameters;3)The layer with the largest comprehensive index of reservoir characteristics in each production layer of the well often forms the dominant water-flooding channel;4)The prediction model is applied on site and the result is highly reliable.