基于融合聚类的篦冷机工况划分与识别研究  

Research on Division and Recognition of Grate Cooler Operating Conditions Based on Fusion Clustering

作　　者：陈梅[1] 陈熙鹏 陈薇[1] 程勇 高翔 褚彪 CHEN Mei;CHEN Xi-peng;CHEN Wei;CHENG Yong;GAO Xiang;CHU Biao(School of Electrical and Automation Engineering,Hefei University of Technology,Hefei 230009,China;Hefei Cement Research&Design Institute Corporation Ltd,Hefei 230009,China)

机构地区：[1]合肥工业大学电气与自动化工程学院,安徽合肥230009 [2]合肥水泥研究设计院有限公司,安徽合肥230051

出　　处：《控制工程》2023年第2期361-367,共7页Control Engineering of China

基　　金：2020年工业互联网创新发展工程-重点领域及特定场景工业互联网平台应用项目(TC200802D);安徽省重点研究与开发计划项目(202104a05020054)。

摘　　要：篦冷机的运行性能是决定整个水泥生产过程的能源利用效率的重要因素,而准确把握工况状态则是保证冷却过程的运行性能处于最优的前提。以篦冷机为研究对象,提出一种基于融合聚类的篦冷机工况划分与评价方法。首先,通过分析与3种典型工况相关的专家知识及操作经验,选取篦下压力和窑主机电流为检测参数;然后,利用融合聚类算法得到不同运行工况下的聚类标签;最后,利用RBF-SVM算法对得到的聚类标签进行训练得到工况识别模型。利用水泥企业的现场运行数据进行了仿真实验,实验结果表明所提方法为提高水泥冷却过程的运行性能提供了一套可行方案。The operating performance of the grate cooler is an important factor which determines the energy efficiency of the entire cement production process. Grasping the working conditions accurately is the prerequisite to ensure that the operating performance of the cooling process is optimal. It takes the grate cooler as the research object, and proposes a grate cooler working condition division and evaluation method based on fusion clustering. Firstly, by analyzing the expert knowledge and operating experience related to the three typical working conditions, the grate pressure and the main kiln current are selected as the detection parameters.Then, the fusion clustering algorithm is used to obtain the clustering labels under different operating conditions.Finally, RBF-SVM algorithm is used to train the obtained cluster labels to establish the working condition recognition model. It uses the field operating data of cement companies to carry out simulation experiments.The experimental results show that the proposed method provides a set of feasible schemes for improving the performance of cement cooling process.

关 键 词：篦冷机 工况识别 融合聚类 支持向量机 

分 类 号：TF355[冶金工程—冶金机械及自动化]