| Gearbox is a common mechanical equipment used to change the speed and transmit power.Due to the poor working environment of the gearbox itself,it is easy to be damaged and fail: the key parts such as gears and bearings have complex processing technology,high assembly accuracy requirements and high failure rate,which is an important reason for machine failure.The normal operation of the gearbox is often accompanied by speed fluctuation.The traditional frequency domain analysis method will appear the phenomenon of spectrum ambiguity that is difficult to identify and it is difficult to diagnose the fault.Even the gearbox that maintains constant speed operation will often have unstable small speed fluctuation over time due to the bad working environment,which not only increases the probability of failure,but also seriously affects the accuracy and reliability of mechanical equipment,And many engineering machinery and equipment are not installed with tachometers,which can not determine the speed in real time and accurately.If the instantaneous speed information can be accurately estimated under the condition of variable speed and lack of tachometer measurement information,and the fault vibration characteristics caused by gearbox fault can be found in time,effective measures can be taken to reduce economic losses and catastrophic accidents.In this paper,the instantaneous speed information in the time-domain vibration signal of gearbox is mined based on the dynamic time warping algorithm.Firstly,the fault vibration signal of gearbox under small variable speed is studied,and then the gearbox fault diagnosis algorithm under small variable speed is proposed.Then,through the research of order resampling technology,the fault diagnosis algorithm of engineering practical gearbox is proposed,Finally,the effectiveness of this fault diagnosis algorithm is verified on two public data sets.The main work can be summarized as follows:(1)the principle of dynamic time warping algorithm is explained,the problems in the application of this algorithm are discussed,and the corresponding solutions are given.Then based on this,a fault diagnosis algorithm under small speed change is proposed,which solves the problem of fuzzy vibration signal spectrum of gearbox without tachometer under small speed change.The effectiveness of this algorithm is verified in the actual experiment of small speed variable gearbox.(2)In view of the difficulty of applying order resampling technology to diagnose the fault of gearbox vibration signal with large speed change without tachometer,a segmented band-pass filter single component extraction algorithm is proposed,and then a single component vibration signal instantaneous frequency estimation calculation method based on dynamic time warping is proposed,which solves the problem that the accurate speed cannot be obtained without tachometer.Finally,the accuracy and effectiveness of segmented bandpass filtering algorithm and speed estimation algorithm based on dynamic time warping are verified by experiments and simulation.(3)Based on the research of instantaneous frequency estimation algorithm and order resampling technology of large variable speed gearbox signal,an engineering practical gearbox fault diagnosis algorithm based on dynamic time warping is proposed,and the order calculation formula of common fault characteristics of gearbox is given.The application ability of engineering practical gearbox fault diagnosis algorithm in practical engineering is verified by two data sets of open gearbox vibration signal. |