Experimental Identification of the Behavior of Glass Fiber Bushing

Two different techniques exist for glass fiber production, direct melt bushing and marble bushing. The purpose of the project was to study the glass fiber production process and investigate the possibilty of constructing mathematical models from experimental data. This part also included implementation of a DAQ system to obtain the expeimental data. A second objective was to construct an efficent temperature control for a producion unit for glass fiber production. The approach taken is to measure the electrical power supplied to the bushing and the glas temperature. From the I/O data a model in the form of a high-order transfer function for the process would be identified. With the model a controller which would minimize process variations would be constructed. The experiments shows that there is a large difference between the two techniques in terms of process disturbances. For the marble melt bushing only approximate models could be constructed. For the direct melt bushing the process disturbances was less important and models with small variances could be constructed. The approach taken would yeild stable, robust controller. However better control could propably be achieved if some measurable process disturbances like the glass temperature at the entry is incorperated. Another approach would be to use an adaptive controller. This approach would be best suited for the direct melt bushing with it's slowly varying process dynamics (Less)