耐火砖全尺寸模糊分类方法及其执行机构动态控制的研究

发布时间：2019-06-27 13:16

【摘要】：耐火砖外形尺寸是否符合公差允许范围是耐火砖质量好坏与否的关键指标之一。理论研究表明，严格控制耐火砖外形尺寸，一定程度上有利于窑炉的砌筑稳定性，且最终影响并决定着窑炉的使用寿命。随着现代科技的发展，在线自动化的检测耐火砖尺寸成为了可能。在现有检测水平的基础上，进一步对耐火砖尺寸进行细化分割并有效聚类，从而建立新的耐火砖模糊分类器模型。更细化合理的分类不仅可以满足产品出口的需求，更重要的是可以减小窑炉砌筑时的尺寸累积误差，延长窑炉使用寿命、减少坍塌风险并使经济效益增加。本文主要针对耐火砖全尺寸模糊分类方法及其执行机构的动态控制进行研究和探讨。研究过程中首先对耐火砖使用寿命的影响机理及各种数学聚类算法及进行了系统的学习和研究，据此在多种聚类分析算法中选定了应用广泛、理论成熟的FCM算法进行聚类分析。第二步通过对比各种现有的聚类算法中所存在的优缺点，分析了FCM算法的局限性，针对FCM需要人为给定聚类数的主观性这一问题，根据聚类有效性的原则，，提出了一种聚类有效性函数与FCM算法相结合，从而自适应的确定最优聚类数的方法，并且通过对已知聚类中心数据的进行Matlab仿真，验证了算法的稳定性和可靠性。第三步为了有效克服算法无法处理符号性数据的性质，对数据提出了处理方案，一定程度上克服了算法的局限性，并通过Matlab仿真分析比对了聚类结果，通过不断增加样本数量的方式考察聚类中心的波动，从而构建了稳定的模糊分类器模型。最后将模糊分类器模型植入PLC控制程序，设计人机界面，建立了人机通讯，并通过程序仿真完成对其执行机构的动态控制。结果表明模糊分类器的聚类结果稳定、客观，其执行机构能准确地完成分类输出。
[Abstract]:Whether the shape and size of refractory brick meets the allowable tolerance range is one of the key indexes for the quality of refractory brick. The theoretical research shows that strictly controlling the shape and size of refractory brick is beneficial to the masonry stability of kiln to a certain extent, and ultimately affects and determines the service life of kiln. With the development of modern science and technology, it is possible to detect the size of refractory brick by online automation. On the basis of the existing detection level, the size of refractory brick is further refined and segmented and clustering effectively, so as to establish a new fuzzy classification model of refractory brick. More detailed and reasonable classification can not only meet the needs of product export, but also reduce the cumulative error of size in kiln masonry, prolong the service life of kiln, reduce the risk of collapse and increase the economic benefit. In this paper, the full size fuzzy classification method of refractory brick and the dynamic control of its actuator are studied and discussed. In the process of research, the influence mechanism of refractory brick service life and various mathematical clustering algorithms are studied and studied systematically. based on this, the widely used and theoretically mature FCM algorithm is selected for clustering analysis. In the second step, by comparing the advantages and disadvantages of various existing clustering algorithms, the limitations of FCM algorithm are analyzed. In order to solve the problem that FCM needs the subjectivity of artificial clustering number, according to the principle of clustering validity, a method of combining clustering validity function with FCM algorithm is proposed to determine the optimal clustering number adaptively, and the Matlab simulation of known clustering center data is carried out. The stability and reliability of the algorithm are verified. In the third step, in order to effectively overcome the fact that the algorithm can not deal with symbolic data, a processing scheme is proposed to overcome the limitations of the algorithm to a certain extent. The clustering results are compared by Matlab simulation analysis, and the fluctuation of clustering center is investigated by increasing the number of samples, thus a stable fuzzy classification model is constructed. Finally, the fuzzy classifier model is embedded into PLC control program, the man-machine interface is designed, and the human-computer communication is established, and the dynamic control of its actuator is completed by program simulation. The results show that the clustering results of fuzzy classifiers are stable and objective, and the actuator can accurately complete the classification output.
【学位授予单位】：辽宁工业大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TQ175.7

【参考文献】