Approach to design tension of filament winding for thick composite pipes using a neural network

A novel design approach was presented to optimizing the winding tension for thick composite pipes in this paper. The theories of filament winding machine control system were introduced, and the factors that influence the forming quality of thick composite pipes were discussed. An approach of calculating the residual inner stress and deformation in tension winding for thick fiber composite pipes according to superposition principle of the elastic theory was advanced. The inner stress distributions in the thick pipes wound with constant tension, constant torque and taper tension patterns separately were compared. A peculiar neural network structure was presented for optimizing the winding tension with error back propagation. Experiments were conducted to identify dominant mechanisms during neural network convergence process. The results show that the winding tension optimized by this network can meet the design requirement with special inner stress distributions such as isostress.