The Parameter Decision And Remote Control For Torsional Pendulum In Sliding Directional Drilling Based On LSTM

In the process of sliding directional drilling,because the drill string is fixed,and it not rotated,the interaction is easy to produce a large friction force between the drill string and the borehole wall,which cause the "supporting pressure".Torque Rocking Technology is commonly used in practical production by applying forward and reverse torque to the drill string.Now,because of the short of relevant theoretical support,it is still necessary to manually adjust the value of torque,which can adjust the tool face to the target value in the process of sliding orientation operation,which depends on the experience of engineers,and the use effect varies greatly among individuals,and the control efficiency is low.In view of the above situation,based on the development project commissioned by Drilling Engineering Technology Research Institute,in this paper a model for predicting the state of tool face based on Long Short Term Memory Network(LSTM)is proposed based on the relevant knowledge in deep learning field,which can make the torsion pendulum system controlled by remote intelligence,mainly including multi-feature fusion processing of sliding directional data,the research on Sliding Directional Intelligent Control Algorithm,analysis of field application examples of the independent algorithm module from three aspects of content.According to the above research content,this paper completes:(1)Multi-system data fusion processing and the construction of operation data set of the drilling field,which lays a good data foundation for the study of Sliding Directional Intelligent Control Algorithm firstly;(2)On this basis,through the analyzing of the sliding directional operation experience of directional engineers,built the state of tool surface prediction model based on LSTM and the model of automatic control algorithm based on the state of tool surface,and docking two sub-modules to form a complete Sliding Directional Intelligent Control Algorithm module,the module has realized the multiple time prediction of tool face in the future,and according to the predicted values,the control parameters of the torsional pendulum system can be given,effectively avoid the limitations of physical modeling and the dependence of engineers' experience;(3)Finally,combined with the framework named Flask,the interaction between the real-time data of the working site and the Sliding Directional Intelligent Control Algorithm module is completed.And we conducted field application tests,the results of tests are analyzed.In the deployment and test of work wells,the result shows that at the scene of the drilling,the Sliding Directional Intelligent Control Algorithm module can accurately predict the state of tool surface,which can control tool surface geared to the target area and be stable gradually,and it can be highly efficient and stable for torsional pendulum system of remote control and data transmission.The desired effect was basically achieved,which is expected to further enhance directional drilling automation and the intelligent level.