Influence of the time-varying gear mesh stiffness on the vibration response of a spur gear pair

• Main Author: Lim, Zheng Ping
• Other Authors: Jaspreet Singh Dhupia
• Format: Final Year Project
• Language: English
• Published: 2014
• Subjects: DRNTU::Engineering::Mechanical engineering
• Online Access: http://hdl.handle.net/10356/60334
• Institution: Nanyang Technological University

Health monitoring is an essential part of preventive maintenance on gearboxes. This process involves the observation of the operation of gearboxes using an array of sensors to pick up important vibration signals which can be further processed to extract information about their health conditions. The main advantage of health monitoring allows the detection of structural faults before it becomes critical. Such preventive maintenance ensures that any faults detected can be corrected to increase the lifespan of a gearbox. This helps to reduce downtime, maintenance cost and occurrence of catastrophic failure which can have safety implications. This report presents a dynamic analysis of a spur gear pair to investigate the vibration response under an analytical model based on the time-varying gear mesh stiffness. The mesh stiffness which is the stiffness excitation of the gear tooth is found to be varying between a high and low value in a periodic manner. A linear crack of varying length at the tooth root set in three crack scenarios are found to cause a deviation of the mesh stiffness. This is a good approximation of a localized gear tooth fault since the gear root is likely the location where the largest stress will be experienced and studies have indicated that most crack propagation paths are rather straight with only a slight curvature. The mesh stiffness is then included in a six degree of freedom model. The degrees of freedoms are two linear displacements and one angular displacement each of the pinion and the gear. The time-domain responses are then obtained by numerical method using the Matlab ODE45 function. Both the responses for a ‘healthy’ and ‘cracked’ gear tooth are compared and found to deviate from each other. The frequency responses of the ‘healthy’ and ‘cracked’ cases are also evaluated to show a clean spectrum dominated by the gear mesh frequency. In the ‘cracked’ case, sidebands, which are spectral components, can be observed around the gear mesh frequency. The frequency spacing of these sidebands is found to be the same as the rotating frequency of the ‘cracked’ pinion, indicating a correlation between the gear fault and sidebands. Statistical measurement of the response shows that the kurtosis indicator performs much better than the RMS in terms of deviation of the healthy values.