多源传感器分层数据融合在煤矿瓦斯预警系统的应用与研究

发布时间：2018-05-06 01:43

本文选题：安全预警 + 多传感器　；参考：《太原理工大学》2014年硕士论文

【摘要】：我国作为资源大国,煤炭在国民生产生活中一直占据不可替代的作用。据统计煤炭在我国能源结构中的占比为74.29%(根据中国统计年鉴)。而煤矿事故频发又成为影响人民安定幸福的巨大隐患,虽然在近年国家的大力投入和科研工作人员的努力下,安全事故总体呈现明显下降趋势,但仍然存在依靠单一传感器预警,预警准确率低,容易出现误报警等现象,既影响生产效率又可能导致煤矿工作人员的麻痹大意。因此本文致力于构建一种综合多种传感器信息,并运用数据融合中分层融合思想的瓦斯预警系统模型。考虑到煤矿灾害的不确定性和非线性问题,本文分析了影响事故因素的内在关联性,提取多种传感器信息。通过对多传感器信息融合一般构架结果分析,提出了由数据层、特征层和决策层组成的瓦斯预警的分层融合结构。为提高多传感器监测系统预警的精度,降低多传感器监测过程中出现的状态不明或状态误判的发生率,具体到每个层次的融合,在数据层,运用层析分析法确定隶属度和相应权数,在特征层运用模糊评价法进行数据融合,在决策层运用D-S证据理论进行数据融合的理论模型。对实验矿井采集到的瓦斯、风速和温度数据进行了数据融合实验。实验结果表明,安全、轻微、危险三种状态隶属度与初始数据比较分别提高8.3%,6%,29.2%,验证了该技术的有效性。
[Abstract]:China as a large country of resources, coal has been playing an irreplaceable role in national production and life. According to statistics, the proportion of coal in China's energy structure is 74.29. And the frequent occurrence of coal mine accidents has become a huge hidden danger affecting the stability and happiness of the people. Although in recent years, thanks to the great investment of the state and the efforts of scientific research workers, the overall trend of safety accidents has shown a marked downward trend. However, there are still some phenomena such as relying on single sensor, low accuracy and false alarm, which not only affect the production efficiency but also cause the workers to be careless. Therefore, this paper is devoted to constructing a gas early warning system model which synthesizes the information of multiple sensors and applies the idea of hierarchical fusion in data fusion. Considering the uncertainty and nonlinearity of coal mine disaster, this paper analyzes the internal relation of the influencing factors of accidents, and extracts the information of various sensors. Based on the analysis of the results of the general framework of multi-sensor information fusion, a hierarchical fusion structure of gas warning is proposed, which is composed of data layer, feature layer and decision layer. In order to improve the accuracy of multi-sensor monitoring system early warning, reduce the occurrence of state uncertainty or state misjudgment in the process of multi-sensor monitoring, specific to each level of fusion, in the data layer, The membership degree and the corresponding weights are determined by the chromatography analysis method. The fuzzy evaluation method is used to fuse the data in the feature layer and the D-S evidence theory is applied to the theoretical model of the data fusion at the decision-making level. The data fusion experiment of gas, wind speed and temperature collected from experimental mine is carried out. The experimental results show that compared with the initial data, the membership degree of safety, slight and dangerous states is increased by 8.3% and 6% respectively, and the effectiveness of the technique is verified.
【学位授予单位】：太原理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TD712

【参考文献】