Study On Optimization Of PDC Bit Based On BP Neural Network

With the increasing use of PDC bits,there have been many new models and new varieties of PDC bits.The existing PDC bit selection methods are mostly based on the experience of using.There is no relation between the formation drilling resistance parameters and the structural parameters of PDC bit,the accuracy of bit selection can not be guaranteed.It is meaningless for creating targeted PDC bit.The relationship between the structural parameters of PDC bit and the drilling parameters of formation is nonlinear.It only has data and records.Therefore,this paper constructs the BP neural network to map the relationship between the two.The structure and learning algorithm of BP neural network are determined through investigation and analysis.The input and output layer of BP neural network is studied and designed based on the analysis of field logging data,geological data and drilling data.The number of hidden layers and the number of nodes of the neural network are determined by the numerical simulation.Construct the BP neural network model.Collect the field data of Qinghai oil field,southwest oil and gas field and Bohai oil field.And set up standard sample set of BP neural network.Based on the BP neural network,the PDC bit optimization software is compiled and do experiment in Linpan area.Field application in Linpan area shows that the drilling rate increased from by 53.47m/h to 69.09m/h,increased by 29.2%,while the drilling rate increased from 19.21m/h to 25.76m/h,increased by 34.1% in the deeper layer.The drill bit recommended by this method match with formation.The footage and drilling rate of bit has been improved significantly.