| Abstract | Parts made from fibre reinforced plastics (FRP) are first laid up on a polystyrene mould. Hence dimensional accuracy is dependent on the quality of the mould. If the parts are oversize, larger than a metre or so, the moulds must be machined in sections and fitted together. There is therefore substantial advantage in a new machining system that can manufacture the mould as one piece. Currently, moulds are machined on a multi-axis robotic or CNC machine that removes material using a milling cutter. When an observation of the current system was conducted on site, the industrial partner gave permission to use their CNC milling robot for cutting expanded polystyrene. A medium density board was also tested. The cutting forces were recorded. It was found that adapting the same cutting strategy was inefficient. The design of such a machine must be based on a new set of requirements.
A number of design factors must be considered. The new machine would be massive and expensive unless three critical issues were resolved at the design stage; cutting approach, instrumentation for control of accuracy, and innovation in the drive.
For removing polystyrene, slicing was considered as an alternative to milling. The fundamental operations of slicing are discussed, from simple pushing of the blade to the incorporation of sideways forces and motions. When using a straight blade, the cutting action can be assumed to act along a straight line. For a curved blade, the straight line was replaced by an effective sector curve, measured from an imaginary centre. While a curved blade and a disk blade work with the same principle, the central axis of a disk blade is fixed, thus the effective sector curve is constant. This allows the disk blade to provide constant sideways force uni-directionally. For slicing, the cutting surface has been evaluated by carving polystyrene with a variety of blades. Later, the force required for cutting with a disk blade was analysed. It was confirmed that a higher cutting speed generated higher cutting forces, but for equal feed rates the cutting force was substantially lower than that required for milling.
The size of the machine was set at five metres by five in plan and two metres in height. It was necessary to devise a position transducer that could operate over these distances, while giving a resolution of a tenth of a millimetre. A transducer was devised based on the visual interpretation of graduations on a precision tape with markings at two millimetre intervals. The concept was embodied using a self-built line-scan camera. The camera performed scans at one or two millisecond intervals, leading to the choice of two millimetre graduations as a compromise between precision and the risk of aliasing at a speed of up to half a metre per second.
A Cartesian gantry configuration was selected as the base platform for the cutting machine. A light weight solution was studied. Traction was provided to each axis using a soft rubber wheel and a pair of DC motors, mounted on the slider. It was possible to combine more than one motor for actuating the slider, if necessitated by heavier loads. In the development of the transducer, a displacement of five metres was achievable with the required placement accuracy. The embedded microcontroller that formed part of the transducer was equipped with an algorithm for precise movement control. Using a nonlinear velocity demand strategy enabled the motion parameters to be optimized.In summary, construction of a full-sized prototype axis has enabled the critical issues to be researched thoroughly. This had created a clear path to the final design. |
|---|