基于信息融合的煤矿采空区火灾预警研究

发布时间：2018-05-17 00:31

本文选题：煤矿采空区 + 数据融合技术　；参考：《西安科技大学》2013年硕士论文

【摘要】：煤炭是我国的重要资源，同时也是一个高危行业，煤矿火灾事故频发，给国民经济、安全都造成了重大损失。煤矿绝大多数的火灾发生在采空区人们不能直视或到达的隐蔽地点，因此煤矿采空区火灾预警的真实性、可靠性和及时性直接影响着企业的生产秩序和企业的经济效益。研究煤矿采空区火灾预警具有重要的现实意义。本文在分析煤矿采空区安全因素的基础上，对各个安全因素与火灾发生的关系进行了较全面的研究，尤其是对煤矿采空区温度与火灾发生的关系进行了深入研究。基于信息融合的煤矿采空区火灾预警主要包括信息采集、信息预处理、信息融合及决策几部分。信息预处理基于小波变换去噪的方法，对带噪信号进行特征提取、低通滤波、重建信号，实现信号的去噪。提出了基于改进的LMBP神经网络技术与D-S证据理论相结合的两级信息融合预测火灾的方法，该方法提取各安全因素的平均值、变化速率值和累积值等特征值，采用改进的LMBP神经网络进行局部信息融合，以构造独立证据理论基本概率分配函数，和基于权值分配的D-S证据理论实现多特征的信息融合判决，实现了煤矿采空区火灾预警，解决了单一安全因素对采空区环境描述较片面的缺陷。利用Matlab平台编写仿真软件，仿真试验结果表明:本文所提出的多特征两级信息融合决策方法能较好的预测煤矿采空区的火灾预警，具有较准确的预警能力和较快的预测速度。本文研究的基于信息融合的煤矿采空区火灾预警经过两级融合模拟仿真试验后，，结果表明可以对采空区的自燃状态进行实时预警，做到防灾、减灾，在现实中具有较高的可靠性和实用性。
[Abstract]:Coal is an important resource in our country, and it is also a high-risk industry. The frequent fire accidents in coal mine have caused great losses to the national economy and safety. Most of the fires in coal mines occur in hidden places where people can not directly look at or reach the goaf, so the authenticity, reliability and timeliness of fire warning in goaf have a direct impact on the production order of enterprises and the economic benefits of enterprises. It is of great practical significance to study the early warning of coal mine goaf fire. Based on the analysis of the safety factors in the goaf of coal mine, the relationship between the safety factors and the occurrence of fire is studied comprehensively, especially the relationship between the temperature of the goaf and the occurrence of fire. The coal mine goaf fire early warning based on information fusion mainly includes information collection, information preprocessing, information fusion and decision making. Based on the wavelet transform de-noising method, the information preprocessing is used to extract the feature of the noisy signal, low pass filter, reconstruct the signal, and realize the signal de-noising. Based on the improved LMBP neural network technology and D-S evidence theory, a two-level information fusion method for fire prediction is proposed. The method extracts the average value of each safety factor, the change rate value and the cumulative value, etc. Based on the improved LMBP neural network, the basic probability distribution function of independent evidence theory is constructed, and the D-S evidence theory based on weight assignment is used to realize multi-feature information fusion decision, and the early warning of coal mine goaf fire is realized. The defect of one-sided description of goaf environment by single safety factor is solved. The simulation software is compiled on the Matlab platform. The simulation results show that the multi-feature two-level information fusion decision method proposed in this paper can predict the fire early warning in the goaf of coal mine, and it has more accurate early-warning ability and faster prediction speed. In this paper, based on information fusion, the fire early warning of goaf in coal mine is studied. After two levels of fusion simulation experiment, the results show that the spontaneous combustion state of goaf can be forewarned in real time, and the disaster prevention and mitigation can be achieved. It has high reliability and practicability in reality.
【学位授予单位】：西安科技大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TD75;TP202

【参考文献】