基于电加热炉温度的预测函数控制系统

发布时间：2018-06-25 02:08

本文选题：电加热炉 + 预测函数控制　；参考：《杭州电子科技大学》2017年硕士论文

【摘要】：温度控制系统是一个大范围工况、大时滞、非线性的过程,但是传统的控制器很难达到工业控制的要求,从而需要研究新的方法去解决这些难题。多模型结构是通过把全局模型分化为一系列的局部模型,利用系统中的局部线性特征可以很好的解决大工况范围造成的非线性的问题。预测函数控制作为一种新型的控制算法,能够根据模型预测未来系统变化的趋势,对未来的变化趋势做出相应的改变,并且这种控制算法的计算量小很适合在实时的控制领域,鲁棒性能强能很好的克服系统的干扰,同时对时间滞后的环节也表现出很好的控制性能。如果把两者的优点相结合起来能够对工业控制的发展起到推进作用。本文主要是针对电加热炉中温度如何快速稳定的达到设定的工作点,而做了以下方面的研究。第一点,针对电加热炉的大工况范围和非线性的特点,把工作范围按照温度的区域进行划分,在局部的工作范围内建立的局部线性模型,从而把非线性系统转为了局部的线性系统。根据划分好的局部线性模型设计相应工作范围内的预测函数控制器,并把Smith预估器引入到控制器的设计中减小了时间滞后的问题,最后当温度跨过不同的工作区间时,利用多模型的切换思想进行控制器间的切换。通过把此方法用在电加热炉的控制,并获得了较好的控制效果。第二点,在对局部模型进行建模时利用整数阶模型并不能反应局部模型的特性,因此提出了一种分数阶模型的拟合。通过分数阶的Oustaloup数值方法把分数阶模型转换为高阶的整数阶模型,从而通过预测函数设计局部模型的状态空间控制器,最后通过各个模型的当前误差确定每个模型的加权系数,从而得到多模型下的控制量。利用此方法在电加热炉中,通过实验的数据分析,表明了多模型的预测函数控制能够很好的处理此类问题。
[Abstract]:The temperature control system is a large scale, time-delay, nonlinear process, but the traditional controller is difficult to meet the requirements of industrial control. Therefore, it is necessary to study new methods to solve these problems. The multi model structure is divided into a series of local models by using the global model, and the local linear features in the system can be used. The predictive function control, as a new control algorithm, can predict the future trend of the system change according to the model, and make the corresponding change in the future change trend according to the model, and the calculation of this control algorithm is very small and is very suitable for the real time control field, and the robust performance is strong. It is good to overcome the interference of the system, and also to show good control performance for the time lag link. If the advantages of both of the two can be combined to promote the development of industrial control. This paper is mainly aimed at the speed and stability of the temperature in the electric heating furnace to achieve the set work point, and do the following research. The first point is to divide the working range according to the region of the temperature and the local linear model set up in the local working range, so as to turn the nonlinear system into a local linear system, and to design the corresponding working range according to the good local linear model. The function controller is predicted, and the Smith predictor is introduced into the controller design to reduce the time lag. In the end, when the temperature crosses the different working intervals, the multi model switching idea is used to switch the controller. By using this method in the control of the electric heating furnace, a better control effect is obtained. Second points are obtained. In modeling the local model, the integer order model can not be used to respond to the characteristics of the local model, so a fractional order model is proposed. The fractional order model is converted to a higher order integer model by the fractional order Oustaloup numerical method, and the state space controller of the local model is designed by the predictive function. Then the weight coefficient of each model is determined by the current error of each model, and the control quantity under the multi model is obtained. By using this method in the electric heating furnace, the experimental data analysis shows that the multi model predictive function control can deal with such problems well.
【学位授予单位】：杭州电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP273

【参考文献】