Axial extrusion-connection technology is one of the important connection technologies for hydraulic piping systems, with high sealing performance and mechanical strength. In this paper, the finite-element-modeling method is used to simulate the experimental process of the connection strength of the axial extrusion joint. The generation mechanism and calculation method of the connection strength are analyzed. To optimize the joint strength, orthogonal testing and grey correlation analysis are used to analyze the influencing factors of joint strength. The key factors affecting joint strength are obtained as the friction coefficient μ1, μ2 between joint components and the groove angle θ1 of the fittings body. The back-propagation (BP) neural-network algorithm is used to establish the connection-strength model of the joint and the genetic algorithm is used to optimize it. The optimal connection strength is 8.237 kN and the optimal combination of influencing factors is 0.2, 0.4 and 76.8°. Compared with the prediction results of the neural-network genetic algorithm, the relative error of the finite-element results is 3.9%, indicating that the method has high accuracy. The results show that the extrusion-based joining process offers significant advantages in the manufacture of high-strength titanium tubular joints.
Keywords: axial extrusion joint; back-propagation neural network; experimental simulation of joint strength; grey relational analysis; joint-strength mechanism.