| The operational status judgment of special vehicles gearbox mainly rely onmanual disassembly, observing. Using regular maintenance and repair methods toprotect its operation, low efficiency of its detection and diagnosis, and even implied agreater security risk. For this reason, the gearbox of special vehicles is taken as theresearch subject, we apply the existing signal processing methods to analyze its signalcharacteristics. Extracting the feature amount that can characterize the runningcharacteristics of cylindrical roller bearings and choosing a suitable pattern recognitionmethod for its fault diagnosis. This approach improves the detection means and theefficiency of diagnosis, provides a practical and feasible means for fault diagnosis ofspecialty vehicles gearbox cylindrical roller bearings, but also has some practicalsignificance.Firstly, this paper conductes experimental study about the specialty vehiclesgearbox cylindrical roller bearings in a self-built bench test bench, analyzes thecharacteristic of each state vibration signals of cylindrical roller bearings collectedfrom the experiments, extractes the wavelet packet energy entropy of each statevibration signals of cylindrical roller bearings, applies support vector machines toidentify failure modes. The main points of the thesis are listed below.1. Based on specialty vehicles gearbox cylindrical roller bearings typical faultstate bench, simulation produce the cylindrical roller bearing four typical bearingfaults pieces, which are rolling elements failure, pitting and indentation, outer slidingwear, normal state. Collecting vibration signal data of cylindrical roller bearings in theconditions of different types of fault condition, also different load torque and differentspeed.2. Theoretically analysis the basic structure and vibration mechanism ofcylindrical roller bearings, study the vibration signal characteristics of cylindricalroller bearings. Qualitative analysis the natural vibration frequency and fault vibrationfrequency of cylindrical roller bearings. Study the different typical state vibrationsignal characteristics of cylindrical roller bearings from the time domain and thefrequency domain.3. Use wavelet packet to decompose rolling failure, pitting and indentation, outersliding wear, normal four states vibration signals of cylindrical roller bearings, then toextract the wavelet packet energy entropy of each frequency band signal as a characteristic quantity of cylindrical roller bearings in different states. Compare thewavelet packet energy entropy of cylindrical roller bearings different status signals.Verify the validity and usefulness of the extracted signal features.4. The support vector machine theory is applied to the special vehicle gearboxcylindrical roller bearing fault diagnosis. The results showed that the wavelet packetenergy entropy and SVM can been effectively used for fault diagnosis of specialtyvehicles gearbox cylindrical roller bearings four typical state, which are rolling failure,pitting and indentation, outer sliding wear, normal state. Also, it gets a higher accuracyrate and can be used for engineering applications, has guidance and practicalsignificance for the fault diagnosis of special vehicles gearbox cylindrical rollerbearings. |
