A Friction Stir Welding machine is proposed and built to allow future research into the process and to provide a framework from which the application of intelligent manufacturing to industrial processes can be investigated. Initially a literature survey was conducted upon which the design of the machine could be based. The conversion of a conventional milling machine into a Friction Stir Welding machine by applying modern monitoring and control systems is then presented. Complete digital control was used to drive actuators and monitor sensors. A wireless chuck mounted monitoring system was implemented, enabling forces, torques, temperature and speed of the tool to be obtained directly from the process. Software based on a hierarchical Open Systems Architectural design, incorporating modularity, interoperability, portability and extensibility is implemented. This experimental setup is used to analyze the Friction Stir Welding process by performing data analysis using statistical methods. Three independent variables (weld speed, spindle speed and plunge depth) were varied and the independent variables (forces, torques, power, temperature, speed, etc) recorded using the implemented software. The statistical analysis includes the analysis of variants, regression analysis and the creation of surface plots. Using these results, certain linguistic rules for process control are proposed. An intelligent controller is designed and discussed, using the derived rules to improve and optimize certain aspects of the process encountered during the experimental phase of the research.
| Identifer | oai:union.ndltd.org:netd.ac.za/oai:union.ndltd.org:nmmu/vital:10823 |
| Date | January 2003 |
| Creators | Kruger, Grant |
| Publisher | Port Elizabeth Technikon, Faculty of Engineering |
| Source Sets | South African National ETD Portal |
| Language | English |
| Detected Language | English |
| Type | Thesis, Masters, MTech (Electrical Engineering) |
| Format | li, 322 leaves, pdf |
| Rights | Nelson Mandela Metropolitan University |
Page generated in 0.0017 seconds