The development of automation and intelligence in geological core drilling is not yet mature.The selection and improvement of drilling parameters are mainly judged through experience,and adjustments to drilling parameters need to be made through the judgment of the core after drilling,which has a certain lag and reduces the efficiency of drilling.Therefore,by building a geological core drilling test bench to obtain drilling da-ta,a Back Propagation algorithm is adopted,with WOB,TOR,Q,and RPM as input and ROP as output.At the same time,the impact of bit wear and bit cutting depth on the experiment is considered.Using each different concrete block as a unit,the training and testing sets are divided into 80/20.After data processing,a total of 6 180 sets of data are obtained for training and testing,and the optimal neural network model is trained to pre-dict the mechanical drilling rate(ROP)with a prediction accuracy of 94.1％.Subsequent-ly,by selecting appropriate drilling parameters,the optimization of geological core drill-ing speed can be achieved.This study provides a reference for the prediction of drilling speed in geological core drilling and the automation of geological core drilling machines.
维普
万方数据