Fault Diagnosis of Sucker-Rod Pumping System Using Support Vector Machine

It is interesting thing to diagnose the faults of Sucker-rod Pumping System by analyzing dynamometer cards, which is inexpensive and easy to obtain. But conventional statistical methods are often ineffective to diagnose the faults of Sucker-rod Pumping System. One reason is that there are more parameters or smaller sample sizes than a simple model in analyzing dynamometer cards, so that their degree of freedom is reduced. Another reason is that the accurate structure of the conventional statistical model diagnosing the faults of Sucker-rod Pumping System is hard to make certain, because of that statistical criteria are crucially dependent on such assumptions as normality, homogeneity, independence. In the paper, we present a SVM-based approach for fault diagnosis of Sucker-rod Pumping System by analyzing dynamometer cards. With the method, we can get the working status of the Sucker-rod Pumping System.