Intelligent Control Research Of Movable Flash Butt Welding Machine In Endless Rolling

Endless rolling technology is a complex system integrated of metallurgical, mechanical, electrical, hydraulic, materials, welding, computer control and other subjects. It uses the welding method to weld the steel head and tail together and makes them continuously go to the billet mill. In the Endless rolling technology the most difficult thing is the connection billets technical problems, that is, you need to link the high-speed steel billets in a very short time. For the large cross-section of welded steel, you can generally use electromagnetic induction welding, laser welding and flash butt welding, etc. At present foreign countries use flash butt welding technology to achieve the billets connection. In this paper, according to the endless rolling technology requirements, some key technologies for movable flash welding machine such as the precision control of the hydraulic systems, the main power circuit design and synchronous control of welding machine walking system are studied further.Precise control of the hydraulic systems is one of the key technologies of endless rolling flash welding. Considering that it is in the flash stage of welding, in order to ensure the smooth progress of welding, the speed of moving clamp is equivalent to the flash speed. The method of self regulation parameters and fuzzy - single neuron network control is introduced. It uses fuzzy technology on single neuron dynamic learning rate adjustment, adjusts learning rate controller on-line, combines fuzzy control and neural control together, makes good use of the advantages of the two kinds of intelligent control methods, so the location of the upset hydraulic system is tracked. When flash welding is in the upset stage, the output of hydraulic cylinders should maintain a consistent and achieve the maximum at the same time, in light of the existing hydraulic machinery coupled parallel simultaneous loading system, we use two fuzzy self-adjustment of PID control strategy, first-class fuzzy PID control for benchmark channel, the second-class for simultaneous access, adjust real-time PID parameters to achieve simultaneous movement. The simulation results show that the control method is effective in reducing the pressure on the two systems synchronous dynamic error.Flash welding machine's power supply system has the characteristics of multiple variables, strong coupling, nonlinear and time-varying, it is difficult to establish their precise mathematical model. Based on the Simulink environment, an integrated model of flash welding power (including the "power converter - digital control system - Dynamic load resistance") is built and the main circuit model and dynamic model of resistance load with non-linear features are built too. Research shows that the simulation model can be built by a real simulation of the dynamic process of the actual system, and provides a theoretical basis for reliable control of flash welding machine's power supply.In the course of flash welder walking process, there is static friction, Coulomb friction and viscous dampers. Walking synchronization system is a typical large inertia, nonlinear systems. According to the flash welding machine for walking synchronous control, the mathematical model of control system is established and a control strategy based on genetic algorithms and BP neural network PID is developed. The method uses the initial connecting weights according to the genetic algorithm optimization BP network and gets a group of combination parameters closed to its initial values, then changes the weights by using the BP algorithm to adjust the PID parameters on line. So the welding machine and the rolling billet are simultaneous. The dynamic tracking performance is improved.