多模型加权自适应控制在中厚板层流冷却系统中的应用

在控制冷却过程中, 中厚板的层流冷却过程厚度总是含有较大的不确定性. 如果用单一的精确的数学模型进行控制, 那么当厚度发生剧烈变化时, 难以获得满意的控制效果. 本文提出了一种基于加权策略的多模型自适应控制方法. 首先根据历史数据分析钢铁企业中不同厚度中厚板的生产过程, 针对生产批次多、生产量大、历史数据丰富的中厚板, 在各厚度工作点处建立多个冷却模型, 并设计相应的PID控制器, 以实现生产批次最多的中厚板终冷温度的精确控制. 对于生产批次少、生产量小导致历史数据匮乏的中厚板, 将终冷温度的实际输出与上述建立的多个冷却模型的输出进行比较, 根据输出偏差计算出各个冷却模型的加权系数, 然后采用多PID控制器加权的自适应方法, 以提高终冷温度的控制精度. 而对于由于新生产计划的下达首次出现的新钢种、新厚度的中厚板, 利用最接近该厚度的加权系数实现初始控制、减少初始误差, 同时实时调整加权系数, 以满足控制精度的要求. 最后通过仿真研究予以验证. 该方法一方面采用PID控制器对生产批量大的中厚板在其厚度附近实现精确控制, 具有较强的鲁棒性. 另一方面, 采用多模型加权方法提高生产批量小或者从未生产过的中厚板的控制精度, 具有一定的自适应能力. 因此, 该方法针对层流冷却过程中厚度高度不确定、快时变的特点, 使控制系统同时具有较强的鲁棒性和自适应能力, 从而实现终冷温度控制精度的提高.

Application of Weighted Multiple Models Adaptive Controller in the Plate Cooling Process

In the controlled cooling process, medium plates usually contain some large scale of uncertainty in the thickness. So, if only one fixed model is utilized constantly, it may not work correctly when the thickness of the plate differs largely from the real one. In this paper, weighted multiple models adaptive controller (WMMAC) is proposed. Multiple models are established and their corresponding controllers are designed according to the characteristics of the thickness of plates under certain circumstance. At each moment, the output error of each model is calculated and used to achieve their weighting coefficients. Finally, a global controller is designed by multiplying the controllers' output with the corresponding weighting coefficients. A simulation result shows that the proposed WMMAC is superior to the conventional controller in the medium plate controlled cooling process and has wide application prospect.