An Improved Negative Selection Algorithm and Its Application in the Fault Diagnosis of Vibrating Screen by Wireless Sensor Networks

摘要

The artificial immune algorithm was applied in the fault diagnosis of vibrating screens in this paper. According to the shortcomings (difficult in obtaining signal samples, hard to explaining fault diagnosis results and lack of continuous learning ability) of traditional vibration diagnosis method, the classical negative selection algorithm was improved and used. By combing of the principles of clonal selection algorithm, the classical negative algorithm was improved to enable it determine fault types properly. And the convergence rate of antibodies generation in the detector was also improved by the optimizing of mutation operators. The fault diagnosis model was then verified by experiments. Due to the experimental conditions, the fault states were simulated by replacing the spring with a smaller stiffness one. The vibrating signals were collected and transferred by a wireless sensor network. The data were analyzed and diagnosed based on the fault diagnosis model. It is proved that the method and the model may be practicable for the fault diagnosis of vibrating screens.