【摘要】 为克服带钢热连轧层流冷却系统中大滞后环节产生的不利影响,提高控制精度,提出了将模糊RBF神经网络与Smith预估器相结合的方法。采用基于改进型模糊C-均值聚类算法的RBF神经网络建立预测模型,获得了较高的精度和较快的学习速度。改进后的模糊C-均值聚类算法具有更好的鲁棒性,且放松了隶属度条件,使得最终聚类结果对预先确定的聚类数目不敏感。将该控制器应用到卷取温度控制中,能把卷取温度控制在698~705℃的范围内,满足了实际生产的需要。更多还原

【Abstract】 In order to overcome the effect from uncertain large time delay,and gain higher precision in laminar cooling system of hot strip rolling mill,a type of Smith predictor based on fuzzy RBF neural network is presented.There are some advantages such as faster speed of study,higher precision by using fuzzy RBF neural network which is based on improved fuzzy C-means clustering algorithm as predictor.The improved fuzzy C-means clustering algorithm has better robustness and makes the cluster results insensitive to the predefined cluster number.The predictor is applied to coiling temperature control with satisfactory results.更多还原