Research On The Construction Method And Application Of Digital Twin For Oil Production Well

As an important part of the development of intelligence,the research on the construction of digital twins in oil production wells has been paid more and more attention.To address the constraints of severe downhole conditions and restricted space for data collection in the process of digital twin construction in oil production wells,a mechanismand data-driven hybrid model is used as the theoretical basis for digital twin construction to comprehensively describe the production process of oil production wells,based on which the prediction and analysis are carried out.A digital twin construction method for oil production wells is proposed in this paper from the industry consensus of the digital twin construction framework and combines the physical process of fluid flow in oil production wells.The integrity and versatility of the digital twin construction for oil production wells are improved with integrated design,and the functions of each module of the digital twin construed are validated in the case study.Based on the basic service architecture of the digital twin,the data acquisition and data lake governance,the construction of the hybrid model,and the provision of cloud platform services of the digital twin for oil production well are studied.The multi-source and multi-type characteristics of the digital twin for oil production wells are considered,and the construction method of the hybrid model based on mechanism and data-driven model is proposed.Based on the data lake and the model,the digital portrait of the oil production well is constructed and the cloud platform services are provided.In view of the production process of oil production wells,according to the degree of difficulty in data acquisition and the complexity of analysis requirements,the hybrid model construction method is proposed in the direction of data flow for four modules of the production process: working condition of lifting equipment,fluid flow status in the well,production performance of oil production wells and the integrated production analysis.During this process,the whole life-cycle production process of the oil production well is abstracted into a mapping relationship among three dimensions: time,space,and parameter,as theoretical support for the function construction of the digital twin for oil production wells.In view of the data characteristics of the digital twin for oil production wells,the whole life cycle of the digital twin for a production well is abstracted into 8 cubic parameter blocks with different data characteristics according to the three dimensions of time,space,and parameters.The data mining of the available data in the production process,and the corresponding parameter mapping relationships of each functional module are studied in detail,and an adapted hybrid model construction method is proposed for each key mapping relationship.Based on the aforementioned theoretical support,case studies are conducted on the constructed functional module of the digital twin for oil production wells in combination with the actual data of oil production wells obtained from oil fields,and the established models are evaluated.In the health management of lifting equipment,56 wells with 4working conditions were identified with the proposed algorithm,and the F1 score over0.8 is reached.In the virtual measurement of production,an error percentage of 3.13%for the gas phase and 8.22% for the liquid phase is achieved,within 10%.In the prediction and evaluation of the production performance of oil production wells,40 wells with a production time of about 1000 days were evaluated,among which,more than 100 abnormal events were predicted in advance.In the integrated production prediction,an accuracy of 98.043% is achieved for 56 wells in 4 formations.The effectiveness of the digital twin construction method for oil production wells is then proved.