The design and implementation of an ultra high cycle fatigue testing machine
McMillan, Tristan John
This work serves to detail the design, implementation and validation of the Ultra High Cycle Fatigue Test rig. The rig has been designed using a modular design methodology. The design philosophy implemented was to keep the overall design as simple as possible so as to reduce manufacturing costs while allowing for modular changes so that the rig can be modified for future conceivable test methods in the ultra-high frequency region. Three versions of the design were created namely the Mark 1, 2 and 3. Each of the versions was an iterative improvement on the previous. The initial Mark 1 was a simple system with no rig and a simple strain measurement system. The Mark 2 consisted of the full test rig setup with both strain gauges and laser acquisition used for the measurement system. Finally the Mark 3 system was created with only the laser used for data acquisition. The final design only implemented the non-contact laser vibrometer measurement system. This was due to the fact that the operational costs were lower with the laser vibrometer, as unlike the strain gauges, it did not need to be replaced after each test. The data was also more repeatable and reliable as the system was stable during testing with no abrupt failures. A 2D finite element method (FEM) code was developed and validated against MSc Nastran for the development of the horn and specimen geometries. The FEM code was used in the design process to accurately predict the geometry such that an axial displacement natural mode is excited at the input frequency of the system. The horn and specimen were designed to resonate at the power generation output frequency of 20 kHz. The final design was validated using comparative data for Aluminium AL2024 with current test data available from literature. This data comparison indicated that the designed rig produces accurate and repeatable data with a good correspondence to similar test data. Furthermore the rig was used to obtain fatigue data for aluminium material AL7075 in the region of 106 to 109 cycles. The data for AL7075 was presented against comparative data as found in the literature survey which indicated that data showed a good correspondence to standard fatigue test data for smooth specimens. Comparisons on the effect of the surface finish on the data obtained using the ultra-high cycle fatigue testing system were conducted and it was found that, as per the literature study, this had a negligible effect on the final result.
This research dissertation is submitted to the Faculty of Engineering and the Built Environment, University of the Witwatersrand, in fulfilment of the requirements for the degree of Masters in Engineering. 26 October 2015, Johannesburg