| As one chief steel production, the media and thick steel plate is widely used, and its demand is more and more. Recent years, with the rapid development of manufacturing industry, the demand of the media and thick steel plate as well as the requirement of the plate quality is needed more and more, especially the precision is needed higher and higher. But the shearing precision is not high at present in home; we have to improve the technology to provide much higher quality plate for the manufacturing industry.In order to improve the shearing precision of the media and thick steel plate as well as the quality of the cutting line, the author researches as follows:Firstly, the author takes the 12-22X2250mm cutting line for example to introduce and study the connection of the main controlling unites and the system's composition of the cutting line.Based on this, mainly on the position control determining the precision, the control method based on fuzzy neural network with a net predictor is designed, which can improve the accuracy of cutting plate comparing to the original fuzzy control. The learning capacity of the neural network will make up the flaws about accuracy in the old control way. The establishment of the fuzzy neural network's model, the improved learning algorithm and the design of the neural network predictor are the emphasized parts of this paper. Through the actual dates and the simulation, the author analyzed the cutting accuracy between the fuzzy control and the fuzzy neural network control. So the conclusion is the fuzzy neural network control designed in this paper has more excellent control characteristics, which satisfy the demand of the high accuracy of cutting production.In addition, the neural network PID control is used in the throwing and stacking process to avoid surface distress and uniform stacking. Using this control, the cutting productivity can be increased and the quality of late can be guaranteed and the automation can be enhanced. By the simulation experiment, the problems in throwing and stacking can be well solved. |
