Daliuta Coal Mine, CHN Energy Shendong Coal Group
Shenyang Research Institute Co., Ltd., China Coal Technology Engineering Group
National Key Laboratory of Coal Mine Disaster Prevention and Control

针对煤矿复杂环境下耦合灾害发生与防治,通过采集不同传感器的多元混合气体(O2、CO、CO2等)以及温度、湿度、大气压力等信息,对数据进行预处理,基于采空区煤自燃与瓦斯灾害信息演化机制,通过数据清洗、数据变换、归一化等数据预处理方法将传感器采集的时序数据转化为类图像矩阵数据,提出了一种新的混合气体识别方法;采用自适应校正方法对数据的分布进行漂移补偿,从分类决策的层面上进行漂移补偿,进而使算法与当前的传感器输出相匹配;完成基于多源时空监测数据的智能预警软件平台开发及系统集成,实现了煤矿采空区热动力灾害预警指标在线监测、智能判识、实时预警,为矿井热动力灾害预测预警工作提供技术支撑。

In view of the occurrence and prevention and governance of coupling disasters under the complex environment of coal mines, the data are preprocessed by collecting the information of multiple mixed gases (O, CO, CO, etc.), as well as temperature, humidity, atmospheric pressure and others from different sensors. Based on the evolutionary mechanism of coal spontaneous combustion and gas disaster information in goaf, the time series data collected by sensors are converted into image like matrix data through data preprocessing methods such as data cleaning, data transformation, normalization and so on, and a new mixed gas recognition method is proposed; Adopting adaptive correction method to compensate for drift in the distribution of data, and performing drift compensation at the level of classification decision-making to match the algorithm with the current sensor output; The development and system integration of an intelligent early warning software platform based on multi-source spatiotemporal monitoring data are completed, achieving online monitoring, intelligent identification, and real-time early warning of goaf thermodynamic disaster early warning indicators in coal mine, providing technical support for the prediction and early warning of thermodynamic disasters in mines.