Study On The Look Ahead And Ultra-deep Azimuthal Electromagnetic Resistivity Logging While Drilling

Given the widespread use of high-angle and horizontal(HA/HZ)wells in production exploration,azimuthal electromagnetic(EM)resistivity logging-while-drilling(LWD)and geosteering technology urgently needs new developments to address the challenge of unconventional reservoir exploration.However,due to the insufficient look ahead and look around detection ability of resistivity LWD,its applications are limited in multi-boundary,far-boundary identification and formation evaluation on site.Therefore,it is necessary to study a new tool design of look ahead and ultra-deep azimuthal EM resistivity LWD for production challenges.In this paper,Fast Hankel Transform(FHT)and Finite Element Method(FEM)methods are proposed for look ahead and ultra-deep azimuthal EM resistivity LWD simulation research.Besides,the forward methods are verified based on the response simulation of Periscope15 tool in different uniform infinite formation models.Second,this paper proposes the preliminary tool design of azimuthal EM LWD based on the analysis of the azimuth signals,voltage amplitude and look ahead distance(LAD)and distance to boundary(DTB)under different tool parameters,environmental factors and instrument structures conditions.Then,this paper quantified the influence of main tool parameters on the voltage and azimuthal detection characteristics by the gray correlation analysis method.The influencing priority analysis of tool parameters is completed at the same time.Through the response simulations of the common EM resistivity LWD tools in different formation models,we found that its look ahead and ultra-deep detection ability is insufficient.Therefore,two tool structures design are studied,including the coils design near the bit and the coils design behind the bit.After reasonably selecting the least associated parameters,the response law of the main influencing parameters is researched.By comparing with the random forest algorithm,the Delaunay triangulation/linear interpolation method is selected in detection data analysis.Then,combined with a large number of numerical simulation results,the azimuthal detection and voltage amplitude of multiple instrument parameter combinations are analyzed.Finally,the EM LWD tool is designed for look ahead and ultra-deep detection and its signals application are discussed for different formations.According to the azimuthal detection characteristics and conventional detection characteristics of five coils combinations in different formation conditions,the applicable conditions of the designed tool are proposed.Due to the surrounding rock,anisotropy etc.influence on azimuthal signals,the formation boundary can be identified difficultly by the response of the designed tool in complex reservoirs.Therefore,the wavelet transform method is studied for boundary identification.An effective and reasonable method for automatic formation division is proposed after the application comparison of five wavelet-based functions in different models,which serves as a reliable theoretical foundation for efficient geosteering and formation evaluation in the future.At the same time,the signals crossplots and the neural network methods are proposed to extract the anisotropy coefficient,horizontal resistivity and tool distance to boundary information.Through the application in the seven-layer model under vertical well and HZ well conditions,we found that the accuracy of crossplots method is lower than that of the neural network method.The neural network method can be applied well in the complicated formation which makes more effective formation information inversion.