Prediction Of Separate Injection Volume Of Water And Design Of Downhole Injection System Based On CNN-LSTM

Stratified water injection is an effective way to improve injection and production structure and enhance oil recovery in the process of oilfield exploitation,which is widely used in the exploitation of high water cut oilfield.In the process of exploitation for oilfield,the requirements of stratified water injection technology are more stringent due to the increasing difficulty of oilfield exploitation.It has become the key of fine and efficient water injection in oilfield to determine the interval injection rate and improve the downhole injection technology.In order to accurately calculate water injection rate in the interval of the injection well,this thesis uses the method of neural network to predict and establishes the layer injection prediction model based on convolutional neural network.The actual production data of the injection well is regarded as the input to train the prediction model,and then the results of the prediction model are compared with the actual layer water injection.For the sake of the prediction accuracy of the prediction model,the model is compared with other water injection calculation methods.In order to further optimize the prediction model,the prediction model of interval injection allocation based on CNN-LSTM is established by combining the model with the long-term and short-term memory neural network.Through the comparison and analysis of the results of the above two prediction models,it is concluded that the interval injection allocation prediction model based on CNNLSTM produces results that are closer and more reliable to the actual results and improve the impact of layered water injection in the oilfield.On the basis of calculating the interval injection rate,the downhole injection strategy is designed in this paper by analyzing the weight value and utility matrix of various factors affecting the layer division and calculating the comprehensive evaluation value of each small layer so as to determine the injection order.Stratified water injection is susceptible to other factors in the actual exploitation.In order to ensure the accuracy of water injection in the underground section,this thesis carries out the comparison and analysis of the common control algorithm so that the incomplete differential PID is selected to adjust the flow of water injection layer.Finally,the control algorithm highlights its effectiveness and superiority in the section flow control by experimental simulation.According to the stratified water injection distribution strategy,this thesis designs the downhole injection distribution system of the oilfield injection well,and then the corresponding interval injection distribution scheme.Finally,the prediction of layer injection rate,priority order of layer distribution and control strategy of injection system are validated and analyzed by analyzing the engineering cases.The experimental results show that the system can effectively regulate the water injection of each underground section and improve the stratified water injection efficiency.