煤矿瓦斯突出灾害动态辨识与预测基础理论研究

Gas outburst is one of the most serious threat to coal mine. The project includes the study on fast dynamic identification and prevention mechanism of gas outburst, explore and solve the scientific problems with new methods and new theories. The object of the project is to establish the model of gas outburst disaster information features of dynamic identification and the model of coal mine gas outburst disaster intelligent forecasting, to solve the key technical problems of dynamic identification and prediction of gas outburst disaster features, and to perfect the theory system of dynamic identification and prediction of gas outburst disaster features. The project breakthrough the key technology in prediction of gas outburst and provide a theoretical basis and technical support for curbing great mine disasters of gas outburst..With the research platform based on topics and the work base of the related research early in the project, we are in-depth study of gas outburst disaster dynamic information fusion and feature extraction method. The key technology of dynamic identification of gas outburst disaster and its implementation method and dynamic forecasting model of gas outburst and prediction mechanism with universal applicability will be studied. Data "mining" and analysis are conducted by monitoring the data resource, a multi-source、multi platform、multi-sensor model of gas outburst in mine disaster for online dynamic identification、hidden danger discriminated will be established. The artificial intelligence method will be used in establishing dynamic prediction model of gas outburst. The basic theories and methods of preventing gas outburst disaster will be provided.

针对煤矿瓦斯突出灾害问题，研究煤矿瓦斯突出灾害的快速动态辨识与预防预测机理问题、探索研究和解决这一科学问题的新方法和新理论。建立瓦斯突出灾害信息特征动态辨识模型和煤矿瓦斯突出灾害的智能化预测模型，解决煤矿瓦斯灾害多特征动态辨识与预测这个关键技术问题，完善瓦斯突出灾害特征动态辨识与预防预测的理论体系，突破煤矿瓦斯突出预防预测的关键技术，为遏制煤矿重特大瓦斯突出事故提供理论依据和技术支撑。.利用课题所依托的研究平台及课题前期相关方面研究的工作基础，深入研究瓦斯突出灾害的动态信息融合与灾害特征提取方法；研究瓦斯突出灾害的动态辨识关键技术及其实现方法；研究瓦斯突出动态预测模型和具有普遍适用性的预测机理。通过对监测数据资源进行数据"挖掘"、分析，建立多源、多平台、多传感器煤矿瓦斯突出灾害在线动态辨识、隐患判别模型，利用人工智能方法建立动态瓦斯突出预测模型，为防止瓦斯突出灾害的发生提供基础理论和方法。

本项目通过理论分析、仿真研究、实验测试、现场运行等研究手段及路线，广泛深入到煤矿企业进行现场调研与测试，获得了大量的煤矿开采地质条件、现有煤矿安全设备设施、现场瓦斯检测信号等第一手材料，在瓦斯突出灾害的动态辨识、特征提取及预测方面进行了深入研究工作。项目进展顺利，按照项目计划任务书进行了相关的研究工作。项目研究了煤与瓦斯突出的检测方法、瓦斯突出辨识系统的特性、瓦斯突出辨识方法、瓦斯突出多源信息融合辨识模型、瓦斯突出耦合辨识模型、瓦斯突出混沌动力学辨识模型、瓦斯突出软测量辨识模型等内容。通过本项目的研究，有效解决多源煤矿瓦斯突出灾害信息的动态特征提取与快速辨识问题，采用数据挖掘方法在线动态提取和辨识瓦斯突出灾害的预兆信息，解决瓦斯突出灾害的动态特征提取和快速辨识这一关键问题；解决我国煤矿瓦斯突出灾害预测基础理论中的关键技术问题，根据我国煤矿井下瓦斯突出灾害发生的复杂性、以及无法建立精确的瓦斯突出灾害变化规律的数学模型等特点。通过对监测数据资源进行数据“挖掘”、分析，建立多源、多平台、多传感器煤矿瓦斯突出灾害在线动态辨识、隐患判别和预测模型，攻克瓦斯监测系统不能及时发现重大瓦斯突出灾害隐患的关键技术。提出利用多种监测信息进行挖掘、分析、处理、融合和综合判断瓦斯突出灾害危险性的先进理论和方法。为发现和找出瓦斯突出预测机理提供理论依据和科学方法。