Drilling Tool Failure Diagnosis Based on GA-SVM

Abstract

Drilling tool failure is a major reason for low drilling speed, thus early drilling tool failure diagnosis is very important. Based on the generalization and approximation of non-linear capability of support vector machine (SVM) and the powerful ability of global optimization of genetic algorithm (GA), this paper establishes combined GA-SVM model through optimizing SVM parameters by GA. Then the model is used to forecast drilling tool failure of actual engineering practice. The experiment results show that optimal parameters searched by GA do significantly improve the forecast performance, and validates the effectiveness of the proposed algorithm, which provides a new method to forecast drilling tool failure.