Research On Oil Well Production Forecast Technology Based On Indicator Diagram

With the proposal and construction of intelligent oilfield ideas,data mining technology has been widely used in the field of oil and gas exploration and development.In data mining methods,machine learning as an important method of data regression and classification has been applied in many fields,mainly to provide new solutions for optimizing the production process of oil fields,such as the measurement and prediction of oil well fluid production,which is now commonly used in China The metering method is to use transportation pipelines to transport the oil well produced fluid to the designated metering station for centralized metering.This method is firstly costly,and secondly,it is also very unfavorable for the development of intelligent and automated management of oilfield metering.Therefore,obtaining a technique with a straightforward measurement process,low measurement process cost and equipped automated and constant metering is an essential task for boosting oil field manufacturing.The indicator diagram liquid measuring technology is currently a relatively stable automated single-well measurement technology as well as has been used extensively in the real output of many oil fields,but there are still some problems.Therefore,this article focuses on the problem of oil well production prediction,using the method of combining the basic theory of reservoir engineering and machine learning to conduct research.The main contents are as follows:(1)Introduced the image meaning of the surface indicator diagram of pumping unit wells and taked correlation analysis method to research the likely combination between the indicator card and oil well production and devised the extraction indicator according to moment feature vector and main component analysis.The eigenvalues of the graph obtains the new eigenvalues of the indicator graph for predicting the production of oil wells.(2)Use machine learning methods to discover the main influencing factors of single well production in actual oil field production.The method of combining quantitative analysis and qualitative analysis is used to screen out the variables used to establish the oil well production prediction model.The qualitative analysis is realized by the decision tree algorithm,and the quantitative analysis is realized by the artificial neural network algorithm.Combining the two analysis results and the basic theory of reservoir engineering,eleven parameters are selected as input variables for establishing the oil well production prediction model.(3)Combining the existing methods of using machine learning methods for output prediction,choose Long Short-term Memory Neural Network(LSTM)as the core algorithm to establish the prediction model,and optimize the research for the shortcomings of the LSTM algorithm.In this paper,the artificial fish school algorithm in the intelligent swarm algorithm is selected to optimize the LSTM algorithm,the basic theory of the artificial fish school algorithm is studied,and the artificial fish school algorithm has slow convergence speed and low optimization accuracy.(4)Established an oil well production prediction model based on the improved artificial fish swarm algorithm and long-short-term memory neural network algorithm,and used Matlab GUI to compile software to realize the oil well production prediction function,and imported the actual production data of the oil field into the software for experimental verification.The verification results show that the prediction model has higher prediction accuracy than the indicator diagram oil measurement technology,and realizes the automatic measurement of oil well production.In this paper,based on the theory of indicator diagram measuring liquid technology,the indicator diagram and related production data of oil Wells are selected as sample data,and the remote and automatic measurement and prediction of oil well output is realized by using machine learning method,so as to further guide oil field production.The prediction precision of the oil well production prediction software based on indicator diagram can meet the precision requirement of single well production measurement,which provides technical support for improving oil field production efficiency and reducing oil field production costs.