Establishment And Application Of Leading Geophysical Model For Shale Gas Horizontal Well Drilling

Accurately establishing the leading geophysical model can use the log data of adjacent wells to predict the log values and geological characteristics of the points on the wellbore track of the shale gas horizontal well to be drilled(such as gamma GR,resistivity RT,density DEN,neutron CNL,organic carbon content TOC,brittleness index BI,gas saturation Sg,etc.),is one of the key technologies for "factory" drilling and geosteering drilling of shale horizontal wells.Among them,the use of logging data to calculate reservoir parameters,and then accurately identify shale gas geology "desserts" and determine the target layer and the guide mark layer are the primary work for establishing a shale gas horizontal well lead geophysical model,LWD(LWD The correction of the environmental impact factors of the curve can not only improve the accuracy of the model,but also provide a basis for the feasibility analysis of the model through curve inversion.This article takes the Weiyuan-Changning Shale Gas Industry Demonstration Zone in China as an example.(1)Using a combination of quantitative and qualitative methods to evaluate the characteristics of changes in LWD curves under different environmental influence factors,construct correction charts,and use a method such as optimal fitting and numerical iteration to establish a model calibration correction model and combine interior pages.Rock core data model feasibility analysis.(2)Based on(while drilling)logging curves,establish a calculation model for organic carbon content,brittleness index,and gas saturation of shale reservoirs,so as to optimize the logging geological parameters that are sensitive to reflect good,medium,and poor shale gas formations.(GR,RT,DEN,CNL,TOC,BI,Sg),combined with test data to construct shale gas sample set,calculate the first and second discriminant vector C from the sample of good and poor shale gas samples.K and the corresponding discriminant function values R1 and R2 establish a good,medium,and poor shale gas layer discrimination plate and a classification feature projection function F.Based on the discriminant plate and projection function values,it is possible to quickly and intuitively identify good and medium wells in a single well section.,Shale gas layer,preferably shale gas geological "sweet spot",and then determine the target layer and the guide mark layer.(3)Analyze the logging response characteristics of different lithology,physical properties,oiliness,and brittle formations along wellbore track locations,and determine the formation dip angle and the distance from the bit to the geological interface in real time based on the LWD and MWD data.The target geological layer(geological "sweets")and the guide mark layer are track-matched to establish the lead geophysical model,and the borehole track-while-drilling log curve-gas reservoir lithology comprehensive mapping technology is used to determine the horizontal wellbore track.Shale reservoir location,extension length,and up and down direction.(4)Using the LM algorithm to optimize the traditional BP neural network algorithm,using the soundings of the neighboring wells,azimuth angles,well inclination angles,and the distance from the drill-to-layer interface as the input neuron information,the shale gas horizontal well test The well response value is used as the output neuron information to form a 5 × 15 × 1 network structure.The LM-BP neural network algorithm is used to train the sample data points,and then the leading geophysical model is established.Compared with the traditional multiple linear regression analysis model,the accuracy of the model is higher.(5)Software programming based on the above research contents(the software includes a total of five modules:the calibration module for environmental influences while drilling curves,the multi-parameter two-vector identification method for shale gas geological "sweets" module,drill bit to geology Interface distance calculation module,predecessor geophysical model module based on track matching method,and predecessor geophysical model module based on LM-BP neural network method).Based on the calibration model of the environmental influencing factors of LWD curves,combined with the(while drilling)calculation method of shale reservoir geological parameters,the target layer of shale gas(geological "sweet spot")and the guide mark layer were optimized,and finally formed a The method of establishing the leading geophysical model,combining programming software such as Fortran,C language,and VB language,programmed the method,and carried out the model feasibility analysis through the logging-in-drilling curve inversion method.The result of field application shows that The method(software)has a good application effect and not only provides a basis for the "factory" drilling model,but also can optimize the borehole track in real time in geosteering drilling and increase the drilling rate.