Nuclear Scale Batching System Parameters On-line Optimization

Nuclear-Scale has played very important role in the material proportioning system. However, mostly in domestic nuclear-scale material proportioning system, we use the control algorithm that uses the constant integral time parameter. We know in factories the controlled objects are various, so it can't reach the optimum performance if we use the same control parameter. Confront with this problem, Mei, ZheJun, who complete his studying work for postgraduate degree last year, designed an off-line optimization way to optimize the parameter of Nuclear-scale material proportioning system. This way has good effect on this problem. But with time passed, the mathematic models of controlled objects may change in some measure, the result of off-line way can't adapt to changed models. So if we don't use this way for a long time, the performance gets worse. Base on Mei's working, it's necessary to find an on-line way that can automatically adapt to changed models. This is the major task of this paper.In this paper, I try to complete an on-line optimization process which used self-adaptive control method in the Nuclear-Scale material proportioning system. I used least squares algorithm to identify the mathematic models of controlled objects, and then, used golden mean algorithm to optimize the integral time parameter. Base on the algorithm research, I have compiled a PLC program to realize the automatic tuning of the control parameter in the Nuclear-Scale material proportioning system. Now the PLC program has been used in the Nuclear-Scale material proportioning system of coking company of Wuhan iron and steel cooperation, and has achieved a good effect. In fact, it's easy to use this program and quick to get the result. Compare with the off-line optimization way, the on-line optimization way can adapt to changed models in a real-time way, and it's more precise because avoiding the errors by manual work.Recently, the research on intelligent optimization algorithms develops fast. It must achieve better performance when applying the new good intelligent optimization algorithm. In this paper, I designed an algorithm combined with genetic algorithm and simulated annealing algorithm to tune the control parameter in Nuclear-Scale system. The simulated result indicated this algorithm has many merits, such as search efficiency and robustness is good. Because of the complicacy of this algorithm, we can't use it on PLC. However, with the rapid development of industry control implements, it may be used in the future.This paper introduces the optimization project in the Nuclear-Scale system ofcoking company of Wuhan iron and steel cooperation detailedly, and gives the comparison of in front of the alteration and after it. Base on the project, I designed a GASA mixed optimization algorithm. And then, the paper gives the simulated result of the GASA mixed optimization algorithm, it proves the superiority of this algorithm.