Dynamic wear identification of elevator’s traction sheave based on multiple acoustic emission information fusion

The wear degree of an elevator’s traction sheave cannot be identified easily, quickly, or accurately using traditional measurement methods. In this study, a new dynamic wear identification method for an elevator’s traction sheave is proposed based on multiple acoustic emission (AE) information fusion. According to the significant differences between the AE signals of running traction sheaves with different wear degrees, AE signals measured by four sensors were collected and analyzed to establish feature databases. The membership functions of each AE signal were optimized based on minimum fuzzy entropy and particle swarm optimization and combined with the fuzziness of AE signals between adjacent wear degrees. Finally, the Dempster-Shafer decision criterion was used to accurately determine the wear degree of the traction sheave. Field experiments were conducted on four randomly selected elevators to verify the accuracy of the proposed identification model. The results showed that the minimum reliability value of the identification result was 0.9945, maximum uncertainty was 0.0001, and identification accuracy was 100%. The proposed identification method for the wear degree of an elevator’s traction sheave could efficiently provide early warnings and improve the safety factor of an elevator.