Evaluation Method Of PDC Drill Bit Rock Drillability For Gravel Stratum In Mahu Area

During the drilling process,the Mahu oilfield has challenges such as difficult bit selection,poor mechanical drilling speed,and protracted drilling cycles in conglomerate formations.Conglomerate formations are distinguished by complicated lithology,gravel growth,and non-homogeneity,which cannot be reliably anticipated by present drillability assessment techniques.In this research,we use indoor experiments to integrate logging data with machine deep learning algorithms to conduct a study on the drillability assessment of PDC bits in the Mahu conglomerate formation,which is important for bit selection and drilling parameter optimization in conglomerate formations.Indoor tests with field conglomerate cores were conducted to evaluate the connection model between conglomerate volume content,median grain size,compressive strength,and rock drillability grade values of PDC bits.Additional logging data(compensation density,sonic time difference,resistivity,and neutron porosity)were used to create a logging assessment model for conglomerate formation drillability.Based on the machine depth learning method,four logging data,namely compensation density,acoustic time difference,resistivity,and neutron porosity,which are closely related to conglomerate rock drillability,are chosen as the input layer of the depth neural network,and the rock drillability grade value of the PDC bit is chosen as the output layer,and a depth neural network prediction model for conglomerate formation drillability is established.Analyzing the activation function,loss function,and network structure parameters yielded the conglomerate drillability depth neural network model,which was generated by training the network using training data.The model was validated using test samples and found to be quite accurate.Based on the aforementioned study findings,Python software for assessing the drillability of conglomerate formations was created.The program was used to assess the drillability of a conglomerate formation in a Mahu well.The results showed that the predicted rock drillability grade values of the PDC bit were in good agreement with the actual mechanical drilling speed variation trend in the field,and the depth learning evaluation method outperformed the logging evaluation method,indicating that it has a good chance of being used in the field.