基于AFSA-FCM的火灾预测与控制系统的研究

发布时间：2018-05-16 05:57

本文选题：煤矿火灾预测 + 模糊C均值聚类　；参考：《辽宁工程技术大学》2017年硕士论文

【摘要】：煤矿火灾是煤炭行业发展的拦路石,对煤矿火灾预测的研究有重大的意义。依照传统预测方法对单一的火灾因素的检测已经不能满足复杂的井下状况,本文对各种因素与火灾的联系进行分析,并分析了火灾的几种燃烧阶段的特点,提出了火灾预测的特征。文中分析了井下火灾预测的控制技术,通过使用井下救灾自动风门技术,在矿井火灾灾变时期可以实现对风门远程控制从而杜绝大量有毒有害气体向其他工作地点侵入和蔓延。文中首先采用模糊C均值聚类算法(FCM)对井下火灾数据进行处理,并将结果与样本进行对比,得出了FCM聚类的正确率,并且对FCM聚类算法应用于井下火灾预测的优缺点进行分析。针对模糊C均值算法的缺点,采用人工鱼群算法(AFSA)对其进行优化,对火灾样本进行预测,得出预测结果和目标函数的收敛过程,并与FCM聚类算法和应用较广的bp神经网络分类算法进行对比,预测准确性较未优化的FCM算法更好,而且不会像bp神经网络一样在小样本数量的情况下,准确率受到过大的影响。从对比中可以看出,AFSA-FCM算法在火灾预测上更有优势,
[Abstract]:Coal mine fire is the block stone for the development of coal industry, it is of great significance to the study of coal mine fire prediction. According to the traditional prediction method, the detection of a single fire factor can not meet the complicated underground conditions. This paper analyses the relationship between various factors and fire, and analyses the characteristics of several combustion stages of the fire, and puts forward the characteristics of the fire. This paper analyzes the characteristics of the fire prediction. In this paper, the control technology of the underground fire prediction is analyzed. By using the automatic ventilation door technology in the downhole disaster relief, the remote control of the air door can be realized in the time of the mine fire disaster to eliminate the intrusion and spread of a large number of poisonous and harmful gases to other working places. To deal with the downhole fire data and compare the results with the sample, the correct rate of FCM clustering is obtained, and the advantages and disadvantages of the FCM clustering algorithm applied to the downhole fire prediction are analyzed. In view of the shortcomings of the fuzzy C mean algorithm, the artificial fish swarm algorithm (AFSA) is used to optimize it, predict the fire samples, and get the prediction. The convergence process of the result and the target function is compared with the FCM clustering algorithm and the widely used BP neural network classification algorithm. The accuracy of the prediction is better than that of the FCM algorithm which is not optimized, and the accuracy rate is not greatly influenced by the small sample size of the BP neural network. The AFSA-FCM algorithm can be seen from the comparison. More advantageous in fire prediction.

【学位授予单位】：辽宁工程技术大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP311.13;TP18;TD752

【参考文献】