| The accuracy of the development dynamic data is an important basis for the dynamic analysis work in the oilfield.Due to a wide range of factors affecting the accuracy of development dynamic data and the inadequacy of data accuracy testing methods and technical systems,the development dynamic data may still have deviations or even errors after the initial inspection,making data quality management more difficult.Therefore,there is an increasingly urgent need for accurate and effective data accuracy analysis for development dynamic data accuracy testing.At present,the main method of oilfield development dynamic data accuracy detection is to adopt the two-level audit mechanism of preliminary rule matching audit and expert audit to ensure the accuracy of development dynamic data,but there are problems such as slow detection speed,low accuracy rate and incomplete mechanism in the real detection process,which makes it difficult to be applied on a large scale.In recent years,spatio-temporal data mining technology has been developing,providing a new way of thinking for intelligent analysis of oilfield data.In this paper,based on the technology of spatio-temporal data mining and the business background of development dynamic data accuracy detection,a method of oilfield development dynamic data accuracy detection based on spatio-temporal data mining is proposed to provide a highly applicable and intelligent service support for the dynamic analysis of oilfield data.This paper provides an in-depth analysis of the actual workflow of development dynamic data accuracy detection and designs a development dynamic data accuracy detection model.Firstly,we analyze the development dynamic data accuracy testing standards and processes,design the general framework of the development dynamic data accuracy testing model,and clarify the research focus on development dynamic data modeling and pre-processing and development dynamic data accuracy testing based on spatio-temporal data mining;secondly,we fill in the missing values and standardize the data after the initial inspection,and then eliminate the dimension to ensure the integrity of the data and provide high-quality development dynamic data for subsequent Secondly,we fill in missing values and standardize the data after the initial check,so as to ensure the integrity of the data and provide high-quality development dynamic data for subsequent testing.By confirming the connected wells and screening the neighboring wells to obtain the range of well groups to be tested,graph theory is introduced,and the spatio-temporal topology of the development dynamic data is modeled by combining the spatio-temporal heterogeneity of the data;again,different weight calculation rules are applied to the modeled data to obtain the spatial correlation of the development dynamic data,and the most effective spatio-temporal relationships of the development dynamic data are filtered and selected to participate in the accuracy testing task by using the weight pruning operation.task;then,by analyzing the characteristics of development dynamic data,based on the design idea of multi-detector combination method based on spatio-temporal mixture mode,the set of development dynamic data accuracy detection algorithms is selected,and the detection of the data is completed by using ARIMA,MGLN and STGCN algorithms,on this basis,GCN and TCN are fused,and a development dynamic data accuracy detection method considering the influence of multiple factors is proposed The purpose of this paper is to enhance the detection effect of methods in certain scenarios,and to analyze the results of multiple detection methods comprehensively based on the combination of assignment and overlap degree;finally,in the context of development dynamic data accuracy detection business,combine the accuracy detection model,design and implement an intelligent detection system for oilfield development dynamic data accuracy,and verify the rationality and effectiveness of this paper’s method through real application effects.The application results show that the development of dynamic data accuracy detection model can efficiently solve the current problems of developing dynamic data accuracy detection,and the developed system can detect data accuracy in a timely and rapid manner,which not only improves work efficiency,but also reduces the subsequent problems such as production plan misclassification and deviation of dynamic analysis work due to inaccurate data,and has high application value. |
