Recognition Of Carbonate Reservoir Types Based On Artificial Intelligence Algorithms

Carbonate oil and gas reservoirs occupies an important position in the world’s oil and gas reserves.At the same time,carbonate reservoirs are also the main reservoir types in related blocks in the Tarim Basin in my country,and they play an important role in my country’s oil and gas energy development strategy.The development of carbonate reservoirs is inseparable from the identification of reservoir types.Therefore,it is of great significance to effectively use existing well test data to obtain reservoir information to the greatest extent to guide the development of this type of oilfield.The application of machine learning methods in the field of well testing has always been the research direction of researchers for many years,and some research results have been achieved.However,in recent years,machine learning has developed into deep learning techniques represented by various types of neural networks.Interdisciplinary research on the identification of carbonate reservoir types remains to be studied.The research focuses on the well test interpretation of carbonate oil and gas reservoirs,combined with the actual data of the mine field,and through deep learning algorithms such as feature value extraction,clustering and neural network training,the types of carbonate oil and gas reservoirs are based on triple-medium permeability.The leakage flow model is divided,and neural network training is further carried out for different types of carbonate reservoirs,so as to achieve the purpose of identifying reservoir types on the algorithm scale.Type identification method of well test interpretation curve for carbonate reservoirs in Tabei area.This paper focuses on the identification of the reservoir type of carbonate oil and gas reservoirs,combined with the actual well test data of the mines in the Tabei area,and uses deep autoencoder feature value extraction and Kmeans++ clustering and other deep learning algorithms to analyze carbonate oil Reservoir types were divided,the data set was expanded using a generative confrontation network model,and neural network training was further carried out for the different types of carbonate reservoirs that were divided,so as to finally achieve the recognition of oil and gas reservoir types at the algorithm level.purpose.Afterwards,the deep neural network,convolutional neural network and long-and short-term memory neural network models were compared and evaluated in the classification of well test curve types in carbonate reservoirs.Finally,the most suitable carbonate in the Tabei area was selected.Reservoir identification method for rock oil reservoirs.This paper studies a set of key techniques for well test curve image processing,such as a feature clustering method based on deep autoencoder technology and an image data set expansion method with a generative countermeasure network as the core.Theories and methods for the identification of carbonate reservoir types based on well curves.Through the comparison of deep neural network,convolutional neural network and long-short-term memory neural network model,it is shown that the convolutional neural network is more suitable for the recognition of the type of carbonate reservoir,and the research results can guide the carbon of the target block.The development of acid rock reservoirs has certain theoretical and practical significance.