Research On Fault Diagnosis Method Of Elevator Guide Shoe Based On CEEMD-TQWT And PSO-LSSVM

Elevators are widely used in high-rise buildings,shopping malls,and company buildings.With the increasing number of elevators,the number of elevator accidents has become more and more frequent,and the safety of elevators has become a problem that can not be ignored.The main parts of the elevator are concentrated in the main body part.The stability and safety of the bodypart are affected by the elevator guide boots.Its configuration and type selection are very important.Once the guide boots do not match the guideway or the installation gap is too large or too small,it can cause the elevator to shake.The research work of this paper mainly includes the following aspects:(1)The vibration signal of the elevator guide shoe is collected by using the acceleration sensor of the PMT elevator comprehensive tester EVA-625,and the three-dimensional acceleration module is fixed on the main body of the vibrator.After the software is set up,the data are collected.(2)The vibration characteristic data of elevator guide boots are extracted.CEEMD can effectively deal with modal confusion after optimization of empirical Mode decomposition method.Several IMF components can be obtained by decomposing the original data.It can solve the problem of incomplete data extraction at low frequency.Then the IMF component is denoised by the adjustable quality TQWT and the IMF component is de-noised as an improved wavelet transform method.The data are processed by adjusting the Q factor.Finally,the data which can represent all the original data features and have less noise can be obtained.(3)The data are classified by support vector and multi-dimension feature is used as the influencing factor.The model is trained and predicted by programming.The support vector machine is suitable for the processing of small sample data.In the case of strong correlation the support vector can determine the type of data by changing the mapping to a higher dimensional space.The kernel function of support vector machine is optimized by particle swarm optimization,and the optimal parameter combination is obtained.The results of LSSVM and PSO-LSSVM are compared.The accuracy of the method is verified.The calculation results show that the accuracy of fault diagnosis in CEEM and TQWT mode achieves the desired effect.After the parameters of support vector machine are optimized by particle swarm optimization,the accuracy of fault diagnosis is improved from 76.7% to 83.1%.