DING Xiaohua;HUANG Kun;YANG Wen;

• 1:NARI Technology Co.
• 2:Ltd.

Abstract：

The cement manufacturing industry has always been one of the high energy consumption industries in China,which is highly dependent on energy. Energy consumption accounts for 40%-60% of production costs. In recent years,according to the " Thirteenth FiveYear Development Guidelines for the Building Materials Industry",cement companies have made great progress in energy conservation.However,compared with the world ' s advanced level,there is still a gap in the comprehensive energy consumption per ton of cement.The cement firing system is the main energy consumption part in the cement production process. The cement firing system carries out complex physical and chemical reactions,involving many links and equipment,and the collected data has the characteristics of nonlinearity,strong coupling,numerous variables and large lag. In recent years,with the development of artificial intelligence and the popularization of industrial data collection,distributed control system( DCS) has been widely used in various industries,and artificial intelligence analysis and optimization methods have become the mainstream of industrial data analysis and optimization. In order to improve the production efficiency of cement production enterprises,and based on the analysis of the historical operating variables and production energy efficiency data of the electric process of the cement clinker burning system,a data-driven energy hybrid algorithm for consumption optimization and parameters recommendation of the cement clinker burning system are used. First,for the parameter selection problem,the average influence value algorithm is used to analyze the energy consumption sensitivity,and the parameters that affect energy consumption are filtered. In the modeling phase,the improved BP neural network is used to model the energy consumption. After obtaining the system energy consumption model,the genetic algorithm is used to optimize it,so that the energy consumption model can be controlled to run with the optimization goal of the lowest power consumption per ton of clinker and can obtain the optimized values of the operating parameters. The algorithm has actually been deployed in the cement clinker firing system of Baishan Cement Plant. The operation results show that the algorithm effectively supports the energy consumption management of the cement clinker firing system. Before optimization,the cement energy consumption is always around 15 000 kWh. Through simulation optimization,the optimal energy consumption is 13 661 kWh,which reduces energy consumption by about 7%. At the same time,the recommended values of characteristic variables can be obtained.

Key Words： cement production;data-driven;optimization of energy consumption;integrated learning;parameter recommendation

Abstract:

Keywords:

Foundation: 国家重点研发计划资助项目(2016YFB0601501)

Authors: DING Xiaohua;HUANG Kun;YANG Wen;

DOI: 10.13226/j.issn.1006-6772.IF20080611

References：

• [1]李承洋,龚先政,周丽玮,等．中国水泥工业节能和减排潜力分析[J]．节能与环保,2019(1):21-25.LI Chengyang,GONG Xianzheng,ZHOU Liwei,et al.Energy conservation and emission reduction potential analysis of China's cement industry[J].Energy Conservation and Environmental Protection,2019(1):21-25．
• [2]刘成基．DCS系统在新型干法水泥生产线中的应用[J]．自动化与仪器仪表,2009(1):67-70.LIU Chengji.Application of DCS system in new dry cement production Line[J].Automation&Instrumentation,2009 (1):67-70．
• [3]WEN Yan,ZHANG Wenming.Study on the evaluating Models of energy consumption in clinker cooler conveying[C]//2009 International Conference on Measuring Technology and Mechatronics Automation.Zhangjiajie:[s.n.],2009．
• [4]赵辉,张宁,蔡万通,等．基于改进多元非线性算法的水泥电耗预测模型[J]．制造业自动化,2015,37(1):74-77.ZHAO Hui,ZHANG Ning,CAI Wantong,et al.Cement power consumption prediction model based on improved multivariate nonlinear algorithm[J].Manufacturing Automation,2015,37(1):74-77．
• [5]LIU Na,JING Shaohong,WANG Xiaohong.Research and application of e-p analysis for energy consumption in cement production[C]//Proceeding of the IEEE International Conference on Automation and Logistics.Zhengzhou:[s.n.],2012．
• [6]王盛慧,赵二卫．基于多元自适应回归样条的水泥能耗评估[J]．化工自动化及仪表,2017,44(12):1130-1134.WANG Shenghui,ZHAO Erwei.Evaluation of cement energy consumption based on multiple adaptive regression splines[J].Chemical Engineering Automation and Instrumentation,2017,44(12):1130-1134．
• [7]王东升,杨友文,田宗军,等．基于神经网络和遗传算法的激光多层熔覆厚纳米陶瓷涂层工艺优化[J]．中国激光,2013,40(9):1-8.WANG Dongsheng,YANG Youwen,TIAN Zongjun,et al.Optimization of laser multilayer cladding thickness nanometer ceramic Coating process based on neural network and genetic algorithm[J].China Laser,2013,40(9):1-8．
• [8]周昊,朱洪波,岑可法．基于人工神经网络和遗传算法的火电厂锅炉实时燃耗优化系统[J]．动力工程,2003,23 (5):2665-2669.ZHOU Hao,ZHU Hongbo,CEN Kefa.A real-time boiler burnup optimization system for thermal power plant based on artificial neural network and genetic algorithm[J].Power Engineering,2003,23(5):2665-2669．
• [9]刘春青,王文汉．基于人工神经网络-遗传算法的展成法球面精密磨削参数优化[J]．工程设计学报,2019,26(4):395-402.LIU Chunqing,WANG Wenhan.Optimization of precision grinding parameters of spherical surface by generation method based on artificial neural network-genetic algorithm[J].Journal of Engineering Design,2019,26(4):395-402．
• [10]费文豪．我国水泥生产工艺与设备的发展分析[J]．工业技术,2011(29):139.FEI Wenhao.Development analysis of cement production technology and equipment in China[J].Industrial Technology,2011(29):139．
• [11]张冬梅．新型干法水泥生产工艺和新设备介绍[J]．硅谷,2008(22):1-12.ZHANG Dongmei.Introduction of new dry-process cement production technology and new equipment[J].Silicon Valley,2008(22):11-12．
• [12]王昭旭,郝晓辰,郭瞳瞳,等．基于深度置信网络的水泥生产能耗时间序列预测模型[C]//第37届中国控制会议．武汉:[s.n.],2018.WANG Zhaoxu,HAO Xiaochen,GUO Zhongtong,et al.A time series prediction model of cement production energy consumption based on deep confidence network[C]//Proceedings of the 37th Chinese Control Conference.Wuhan:[s.n.],2018．
• [13]李敏强,徐博艺,寇纪淞．遗传算法与神经网络的结合[J]．系统工程理论与实际,1999(2):65-69.LI Minqiang,XU Boyi,KOU Jisong.Combination of genetic algorithm and neural networks[J].Systems Engineering Theory and Practice,1999(2):65-69．
• [14]翟宜峰,李鸿雁,刘寒冰,等．用遗传算法优化神经网络初始权值的方法[J]．吉林大学学报,2003,33(2):45-50.ZHAI Yifeng,LI Hongyan,LIU Hanbing,et al.A method for optimizing the initial weights of neural networks with genetic algorithm[J].Journal of Jilin University,2003,33(2):45-50．
• [15]李扬．基于BP神经网络的钢材力学性能预测算法[J]．数字技术与应用,2013,13(2):45-67.LI Yang.Prediction algorithm for steel mechanical properties based on BP neural network[J].Digital Technology and Application,2013,13(2):45-67．
• [16]高倩．基于大数据的能耗特性分析方法研究[D]．保定:华北电力大学,2017.GAO Qian.Analysis of energy consumption characteristics based on big data[D].Baoding:North China Electric Power University,2017．
• [17]陈荣武,刘莉,郭进．基于遗传算法的列车运行能耗优化算法[J]．交通运输工程学报,2012,12(1):108-114.CHEN Rongwu,LIU Li,GUO Jin.Optimization algorithm of train operation energy consumption based on Genetic Algorithm[J].Journal of Transportation Engineering,2012,12(1):108-114．
• [18]马永杰,云文霞．遗传算法研究进展[J]．计算机应用研究,2012,29(4):1201-1206.MA Yongjie,YUN Wenxia.Research progress of genetic algorithm[J].Computer Application Research,2012,29(4):1201-1206．
• [19]闵惜琳,刘国华．人工神经网络结合遗传算法在建模和优化中的应用[J]．计算机应用研究,2002(1):79-80.MIN Xilin,LIU Guohua.Application of artificial neural network combined with genetic algorithm in modeling and optimization[J].Computer Application Research,2002 (1):79-80．
• [20]李超．基于自适应遗传算法的BP神经网络预测研究及应用[D]．太原:山西师范大学,2012.LI Chao.Prediction research and application of BP neural network based on adaptive genetic algorithm[D].Taiyuan:Shanxi Normal University,2012.