Research On Elevator Fault Diagnosis Method Based On Vibration Feature Extraction

Recently, with the rapid development of national economy, more and more high buildings have been built. The trade of elevator is embodying its indispensible value with the result of its wide application. However, with the wide application in all kinds of supermarkets, dwellings and company buildings, more and more problems relative with elevators are constantly exposed in the front of the residents. There are data showing that the total number of domestic elevators has surpassed four million sets by the end of the year 2015. Compared with last year, the growth rate has increased by another 8% to 20%. With the increase of total elevator, more and more elevator accidents occur. Although the technology of manufacture and check has been greatly promoted, the number of elevator accidents still increases. Therefore, it has been the focus of modern society to improve the safety and comfort of elevators as well as to predict elevator malfunction accurately and efficiently. In the meantime, these problems have become the emphasis and difficulty supposed to be overcome by elevator trade.The research work of the essay mainly includes the following aspects:(1) To explore the relationship between the signal of accelerated velocity of elevator vibration and the malfunction of the elevator guide shoe on the basis of analysis of vibration mechanism of elevator car. This research proceeds with four conditions of elevators, including normality, tightened guide feet both on the left and right, spring-loosed guide foot and severely-rubbed foot liner. This research not only has important research value in the safety of using elevators, but also provides quantification standard of malfunction traits for the malfunction check of elevators.(2) The collection of vibration signal of elevators. The essay adopts EVA-625 type elevator quality detector, which is provided by the special equipment security detection research institute of Yunnan. This system can precisely quantify the metrical data of accelerated velocity and noise, which lays the foundation for the follow-on research.(3) The extract of trait parameters of elevator vibration signal. The essay starts with the time-domain analysis on the elevator vibration accelerated velocity signal and makes comparisons of time-domain index of elevator vibration signal among four working conditions. Then, with the application of wave decomposition, the essay conducts the analysis of leptokurtosis index on the original vibration signal and decomposed signal of all levels. Finally, with the application of wave decomposition, energy value of all nodes of elevator vibration signal under four operation conditions is acquired. With that, the trait vector reflecting the information of electric malfunction can be mapped.(4) The diagnosis of elevator malfunction. Three methods are adopted to optimize the parameters on the basis of LSSVM, which are K-CV-LSSVM, GA-LSSVM and PSO-LSSVM. It is found by means of comparison that the recognition rate of PSO-LSSVM is closet to the accuracy and can classify the elevator malfunction successfully, which proves its availability.