回转窑煅烧工艺参数配置优化算法研究

发布时间：2018-04-19 03:16

本文选题：回转窑 + K-means　；参考：《北方工业大学》2017年硕士论文

【摘要】：铝电解行业的发展与中国经济的健康发展息息相关,本论文通过研究铝电解生产过程中各个阶段的工艺流程,旨在提高生产效率、减少资源浪费、促进安全生产。在铝电解行业中,回转窑是对石油焦进行一系列复杂处理的大型环状设备。通过研究回转窑煅烧过程的工艺参数,运用数据挖掘理论,分析工艺参数对煅后焦质量参数的影响,为科学认识回转窑工艺参数配置提供帮助。论文研究了铝电解回转窑煅烧工艺的理论知识及相关技术,包括回转窑煅烧工艺流程、回转窑质量参数评定标准,学习软件工程相关理论,学习数据挖掘相关知识,研究课题需要的相关算法,包括K-means聚类算法、主成分分析法、BP神经网络算法、粒子群优化算法等现代智能领域相关算法。具体研究内容包括:首先,本文研究回转窑煅烧工艺的理论知识及相关技术,对回转窑质量参数进行分析,选用煅后焦的粉末电阻率和真密度作为分析对象,研究数据挖掘相关理论,运用K-means聚类分析算法对回转窑煅烧质量参数进行聚类分析,根据质量参数数据与工艺参数数据时间对应的特点,将K-means算法分析后质量参数的分类与工艺参数的分类对应,采用主成分分析法对工艺参数训练样本和测试样本降维。其次,本文针对粒子群算法容易陷入局部极值的问题,提出基于邻代竞争的杂交粒子群优化算法,通过融合BP神经网络挖掘算法,演化成DCHPSO-BP算法,改善收敛效率低和容易陷进个体极值的缺陷,然后对降维后的工艺参数样本进行训练学习,预测回转窑工艺参数对质量参数的影响,构建回转窑配置参数预测模型。最后,运用软件工程的设计思想,采用SSH集成架构,构建完整的、轻量级的J2EE软件开发模型,设计并实现了回转窑煅后焦质量分析系统。系统设计的目的是通过配置工艺参数,使用数据挖掘方法预测出该工艺参数配置会生产出何种类型的煅后焦,并且为企业生产决策提供帮助。
[Abstract]:The development of aluminum electrolysis industry is closely related to the healthy development of Chinese economy. This paper aims to improve the production efficiency, reduce the waste of resources and promote safe production by studying the technological process of each stage of aluminum electrolysis production.In aluminum electrolysis industry, rotary kiln is a series of complex treatment of petroleum coke ring equipment.By studying the process parameters of rotary kiln calcination process and using data mining theory, the influence of process parameters on calcined coke quality parameters is analyzed, which provides help for scientific understanding of technological parameters configuration of rotary kiln.In this paper, the theoretical knowledge and related technology of aluminum electrolysis rotary kiln calcination process are studied, including the process flow of rotary kiln calcination, the evaluation standard of rotary kiln quality parameters, the relevant theory of software engineering, and the relevant knowledge of data mining.The related algorithms needed in this paper include K-means clustering algorithm, principal component analysis (PCA) algorithm, BP neural network algorithm, particle swarm optimization (PSO) algorithm and so on.The specific research contents include: firstly, this paper studies the theoretical knowledge and related technology of rotary kiln calcination process, analyzes the quality parameters of rotary kiln, and selects the powder resistivity and true density of calcined coke as the analysis object.The related theory of data mining is studied, and the K-means clustering analysis algorithm is used to analyze the calcination quality parameters of rotary kiln. According to the characteristics of the time correspondence between the quality parameters data and the process parameters data,The classification of quality parameters after analysis by K-means algorithm corresponds to the classification of process parameters, and the dimension reduction of process parameters training samples and test samples is adopted by principal component analysis.Secondly, aiming at the problem that particle swarm optimization is easy to fall into local extremum, a hybrid particle swarm optimization algorithm based on neighbor competition is proposed in this paper. By combining BP neural network mining algorithm, the hybrid particle swarm optimization algorithm evolves into DCHPSO-BP algorithm.The defects of low convergence efficiency and easy trapping into individual extremum are improved, then the sample of process parameters after dimension reduction is trained and studied, the influence of process parameters on quality parameters of rotary kiln is predicted, and the prediction model of configuration parameters of rotary kiln is constructed.Finally, using the design idea of software engineering and SSH integrated framework, a complete and lightweight J2EE software development model is constructed, and the quality analysis system of calcined coke in rotary kiln is designed and implemented.The purpose of the system design is to use the data mining method to predict what type of calcined coke will be produced by configuring the process parameters and to provide help for the production decision of the enterprise.
【学位授予单位】：北方工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TF351;TP18

【参考文献】