Inversion of Azimuthal Electromagnetic Wave Resistivity LWD Based on Deep Residual Network
Azimuthal electromagnetic wave resistivity logging while drilling(LWD)can provide abundant subsurface information and help to determine reservoir location and complete boundary detection.However,the common iterative inversion method based on physical equations has low computational efficiency and is limited in real-time geosteering.Therefore,an intelligent inversion method of azimuthal electromagnetic wave resistivity logging data based on a deep residual network(ResNet)was proposed.The method replaced the convolution and pooling layers in the residual block with fully connected layers and used a multi-head attention mechanism to understand the relevance of the input data,so as to solve the nonlinear regression problem.By evaluating the depth and width of the model and using a Bayesian optimization tuning algorithm to find the optimal hyperparameters of the electromagnetic wave resistivity inversion method,the performance of the inversion model was improved.The method showed good accuracy in model experiments,with an average accuracy of 98.5% .In the actual logging data,the average accuracy rate was 97.2%,and the single point inversion time was about 0.01 s.Intelligent inversion method of azimuthal electromagnetic wave resistivity logging data could quickly and accurately invert the azimuthal electromagnetic wave resistivity logging data.
deep residual networklogging while drillingazimuthal resistivitydeep learningmulti-head attention mechanisminversion
万方数据
维普
