Research On Reservoir Geological Model Revising Method And Application Based On LWD Information

Reservoir geological model updating,as the foundation of drilling and development,and as a bridge and bond between the two,has been widely studied by scholars at home and abroad.Many scholars have made a lot of meaningful explorations in lithology identification based on logging while drilling(LWD)and reservoir geological model updating at different development stages,but the related research results of real-time updating of geological model with both fast and effective(meet the actual engineering needs)are rarely reported.First,because LWD is mainly used to guide geosteering operations,it is usually used to effectively identify the lithology of reservoirs while drilling.It is difficult to interpret the properties characteristics of reservoirs while drilling,which makes it impossible to provide accurate data basis for real-time updating of reservoir geological models.Second,although mature commercial geological modeling software has realized the effective combination of LWD and software,due to technical constraints,the use of the software is limited,and the logging interpretation method of the software is relatively traditional,which makes it difficult to realize real-time data transmission and accurate model establishment.Third,due to the existence of LWD device "zero length"(the distance between the logging tool and the bit),the reservoir characteristics at the horizontal well bit cannot be accurately interpreted in real time and used for model updating,which makes the model updating lag.These three aspects seriously hinder the development of reservoir geological model updating technology,making it difficult for people to gradually understand reservoirs as drilling progresses,limiting the effective combination of drilling engineering and development engineering in real time.Based on LWD technology,this paper uses machine learning,logging interpretation,computer science,reservoir geological modeling and other interdisciplinary theories and methods to conduct the following researches in order:reservoir characteristics interpretion while drilling,real-time geological model updation around current well,and the lithology correction of horizontal well bit real-time.Firstly,according to field actual data filtering logging sequences,based on effective logging data,rely on the Python language,and combined with machine learning method of support vector machine(SVM),random forests(RF)and gradient tree(GBDT)algorithms,established interpretation models of reservoir lithology,porosity,permeability characteristics and oil-gas-water layers.The parameters of each model are optimized through cross validation and LWD data are interpreted in real time by interpretation models.Secondly,based on the convenience of the Ocean secondary development platform and the functionality of Petrel software,the real-time transmission plug-in for the current well trajectory and reservoir property interpretation results is compiled by using C SHARP language,and the automatic update module for geological model is established.Finally,based on the established measurement point and vertical reservoir boundary distance(D-MP-VRB)database,D-MP-VRB prediction models are established using SVM,RF,neural network(NN),and extreme gradient boosting tree(XGBoost)algorithms,respectively.The parameters of each model are optimized through cross validation,and the prediction formula of drill bit and vertical reservoir boundary distance is established.Combined with case studies,the following main conclusions are obtained:(1)For real-time interpretation of reservoir characteristics while drilling,based on well logging data in Yan'an gas field,taking into account the interpretation accuracy and training time of model,respectively after 262 times,1094 times and 119 times experiments chose RF algorithm to establish reservoir lithology interpretation model,SVM algorithm to establish interpretation models of porosity,permeability and oil-gas-water layers,realizing accurate real-time interpretation of reservoir characteristics while drilling;(2)For the real-time update of the geological model around the current well,based on a simple study aera,the real-time transmission plug-in is used to realize the seamless connection between the logging interpretation results and the geological modeling software,and the automatic update module is used to realize the automatic update of the geological model around the current well area with any length as the distance step;(3)For the real-time correction of the lithology at the drill bit of the horizontal well,based on the logging data of Changqing Oilfield,through 1320 experiments,the D-MP-VRB prediction model established by XGBoost algorithm to predict the step length of 2m category,the effect is better.The distance prediction between the horizontal well bit and the vertical reservoir boundary is realized,real-time lithology correction at the bit is realized,and the adverse effect of “zero length” on the lithology prediction at the bit is reduced;(4)A set of reservoir geological model re-correction method based on LWD information is formed,and the workflow of the method is proposed.It is applied to the Sulige gas field and the reliability is demonstrated by the posterior method.Through probability distribution consistency and adjacent well inspection method,verified the accuracy of the updated geological model.Realized the reservoir lithology correction at the bit in real time,and proved the feasibility and versatility of the method.The research results of this paper provide new method and technical mean for the real-time modification and update of the reservoir geological model,and realize the accurate update of the geological model while drilling.Thereby realize the continuous improvement of the reservoir geological model during the drilling process,which facilitate the effective combination of drilling and development work in real time,and provide an important guarantee for the construction of bridge and bond between drilling engineering and development engineering.