| The calculation of controlled reserves in shale oil wells is an important means of evaluating the exploration value and development potential of shale oil,which is of great significance for formulating reasonable exploration and development plans and decision-making.Reservoir parameter prediction and reserve calculation are two important steps in oil and gas exploration and development.The accuracy of reservoir parameter prediction determines the accuracy and reliability of reserve calculation.This paper first presents a spatial XGBoost interpolation method fused with semi-variogram functions for predicting shale reservoir parameters.Finally,by combining the prediction results of reservoir parameters,a well controlled shale oil reserve calculation method is proposed based on the Monte Carlo method.The specific research content is as follows:This paper proposes a spatial XGBoost interpolation method fused with semi-variogram functions.This method is based on machine learning and Kriging interpolation methods to interpolation prediction of shale reservoir parameters.By comparing multiple interpolation methods,the spatial XGBoost interpolation method fused with semi-variogram improved the accuracy by more than 15% compared to traditional methods and by more than 4% compared to conventional machine learning spatial interpolation methods.This method not only considers the advantages of Kriging interpolation,but also combines the characteristics of machine learning algorithms,allowing for more accurate spatial interpolation prediction of data.This paper proposes a well controlled shale oil reserve calculation method based on Monte Carlo method,which improves the accuracy and reliability of shale oil well controlled reserve calculation.Firstly,I divide the calculation area into units and use known data to determine the probability distribution function of reservoir parameters in a part of the unit grids.Then,Monte Carlo method is used to randomly sample the parameters of units with known parameter distributions,and the corresponding parameters of units with unknown parameter distributions are obtained through the proposed spatial interpolation method.Finally,the volume method is used to calculate the reserves of each unit,and the well controlled reserves of the area are obtained by accumulating the calculation results of all units.Due to the use of Monte Carlo method for sampling and calculating reserves,the final reserve results are presented in probability form,making the calculation results more reliable. |
