Analytic Technique of Drillstem Failure Based on Support Vector Machine Technology and Clustering Theory

An analyzed model of drillstem failure reason has been constructed based on SVM technology and clustering analysis theory in this paper, and the optimal kernel function and several appropriate factors also have been obtained by training the sample data. Using the drillstem failure data come from Daqing oilfield, according to the requirements of the model, the data have been pretreated and analyzed, finally the essential reasons for the drillstem failure here have been obtained successfully; what´s more, six precautionary measures have been pointed out to solve the very problems in this paper. As showing in the paper, the new model is realer and more reliable to calculate and analyze the drillstem failure in drilling engineering, and it has been proved to provide a new technique to enrich the basic theory of drilling engineering.