The Research Of The Setting Parameters Of Magnesium Thin Strip By Twin-roll Casting

Twin roll strip casting technology is currently the most popular and potential technology which has been realized in the laboratory. Some developed countries not only have taken the lead in the research of twin roll casting technology, but also have been the first to industrialized produce. However, our country’s development speed is slower than the developed countries’, the research of this technology is still in the stage of laboratory production.Liquid metal, the raw material, is poured into the space between the opposite rotating casting roll which is regarded as crystallizer. It is a whole complete production process of metal thin strip with liquid metal. It makes casting and rolling to meet on the same equipment. And it has a lot of advantages, such as less process,simpler production equipment, lower production costs, more energy saving, compared with the traditional process. Therefore, the research of this technology is very important to the industrial alloy strip production.In the twin roll strip casting process, there are many parameters that affect the quality of the strip, such as pouring temperature, cooling water flow rate. One of the most important parameters is the location of the kissing point which is directly affects the quality of the strip. As we know, there are many parameters that affect the location of the kissing point, and the most important parameters are casting rolling force and roll speed. In this paper, to establish math model respectively for the force of casting rolling, the speed of roll and the location of kissing point, and to derive the relationship between the location of kissing point and the other two parameters.Building the location controlling model of kissing point, based on technological characteristics, we can get the presetting value of casting rolling force and roll speed.Taking the location of kissing point as the control object, and real-time monitoring the figure of casting rolling force, it can makes the location of kissing point stabilized at the target height by adapting PID for systematic control. In order to improve the response speed, we need to refer to the method of the parameters self-learning. In this approach, it will improve the efficiency by establishing short-term self-learning models for the presetting values of casting rolling force and rolling speed, and it also will makes the system adjusting them next time under the same condition.Due to the influence of some parameters, such as uncontrollable and non-measurable parameters, environmental interference and system error in the process of twin roll strip casting, the actual strip casting production system is a nonlinear system, and it can not fit the method of traditional PID control. So establishing a PID controller model which based on neural network by combining the advantages of neural network and PID controller can makes the system effectively controlled. The so-called PID controller based on neural network adjusts the parameters of the PID controller according to the running state of the system, to achieve the optimization of some certain performance index. That means,pk,ik,dk,the three adjustable parameters of the PID controller, are relatively to the output of the output layer neurons in the neural network.In this paper, the experimental research platform is based on the Magnesium Casting Engineering Laboratory of University of Science and Technology Liaoning.We not only introduce the hardware equipment in this laboratory and the software of human-computer interaction, but also analysis the influence of some process parameters on the results of the experiment. It can validate its feasibility and superiority by comparing the MATLAB simulation results of the PID control based on neural network and the traditional one.