煤矿生产安全时序数据预测方法研究

发布时间：2018-05-24 21:27

本文选题：多传感器 + 融合　；参考：《山西大学》2016年硕士论文

【摘要】：我国产煤量世界第一,并且煤在日常所需能源消耗中约占70%左右,然而矿工死亡总量是其他国家的3倍,所以煤矿作业安全就成为了我们重点解决的困难。并且在日常生产中,瓦斯事故是威胁安全作业的首要因素,又因井下环境变幻莫测,瓦斯浓度受多方环境因素影响,比如温度,风速,一氧化碳等,所以如何对煤矿井下瓦斯浓度以及其它时序数据的实时监测成为了煤矿安全生产过程中迫切需要解决的问题之一。本文首先对目前国内外在煤矿安全生产预警领域中所用到的预测方法的现状进行综述,在此基础上,解析了目前存在实时监测系统的不足,通过对数据融合技术、单一或组合预测方法的研究,针对了单传感器预测数据源单一存在片面性问题以及单一预测数据模型精度低和未曾同时考虑到数据的线性和非线性情况,实现了矿井下数据预测系统。1、利用数据融合技术对瓦斯浓度,温度,风速三种数据进行融合计算;2、针对数据融合后的时序数据线性情况,采用指数平滑预测模型进行单一预测,得出实验结果进行误差分析;3、针对数据存在的非线性情况,采用独立的RBF预测算法,获取数据完成误差分析;4、将两种作用因素均涉及到,利用组合算法实现实验第三部分,完成误差分析;5、针对三种煤矿井下时序数据预测方法的研究,实现最终系统,分析预测结果,得出最佳方案,进而能为煤矿作业过程中的决策提供依据。
[Abstract]:China has the largest coal production in the world, and coal accounts for about 70% of the daily energy consumption. However, the total number of mine deaths is three times that of other countries, so the safety of coal mining has become our key problem. And in daily production, gas accident is the primary factor threatening safe operation, and gas concentration is affected by various environmental factors, such as temperature, wind speed, carbon monoxide, etc., because of the unpredictable underground environment. So how to monitor the gas concentration and other time series data in real time has become one of the urgent problems in the process of coal mine safety production. This paper first summarizes the present situation of prediction methods used in the field of coal mine safety production warning at home and abroad, on the basis of this, analyzes the shortcomings of the current real-time monitoring system, through the data fusion technology, The research of single or combined prediction method aims at the one-sided problem of single sensor prediction data source and the low precision of single prediction data model and the fact that the linear and nonlinear data are not taken into account at the same time. The data prediction system of coal mine is realized. The data fusion technology is used to calculate the gas concentration, temperature and wind speed. According to the linear situation of the time series data after data fusion, the exponential smoothing prediction model is used to carry out a single prediction. According to the nonlinear situation of the data, the independent RBF prediction algorithm is used to obtain the data to complete the error analysis. The two kinds of action factors are involved, and the third part of the experiment is realized by using the combination algorithm. The error analysis is completed, aiming at the research of three kinds of methods of time series data prediction in coal mine, the final system is realized, the forecast result is analyzed, and the best scheme is obtained, which can provide the basis for the decision making in the process of coal mine operation.
【学位授予单位】：山西大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TD76;TP202

【相似文献】