Support vector machines based approach for fault diagnosis of valves in reciprocating pumps

Abstract

Support vector machines (SVMs) represent an approach to pattern classification. The paper presents a SVMs based approach for fault diagnosis of valves in three-cylinder reciprocating pumps. The vibration signals collected from pumps are preprocessed with the wavelet packet transform and time-frequency information is extracted as the character vector for training mid testing the SVMs. To classify multiple fault modes of valves, a SVMs based multi-class classifier is constructed and used in the valve faults diagnosis. The results in experiments show that fault types and positions of faulty valves can be identified and diagnosed by the above method. Furthermore, compared with the results of a BP network, more excellent diagnosis accuracy indicates the potential of the SVMs techniques in machinery fault detection.