基于数据挖掘的煤矿安全监测系统研究-控制理论与控制工程专业论文.docx

摘 要 本文对数据挖掘技术在煤矿安全监测系统中的应用作了初步的研究与探讨。同 时针对目前煤矿安全监测系统中存在的不足，提出了改进的煤矿安全监测方案，基 于 CAN 总线与无线射频技术相结合的煤矿安全监测系统。实现了部分硬件的设计。 该方法弥补了传统有线传输只能测量固定地方的瓦斯浓度，缺乏移动性的问题，又 解决了无线传感器网络的电源供给有限的弊端，综合了两种方法的优点，可以更好 的进行监测。 在软件方面，首先提出应用数据挖掘的模糊聚类方法和模糊评判技术，进行井 下环境监测区域的划分和危险等级的判定，提出了综合距离度量和相似度量的新统 计量，MATLAB 仿真实验结果表明，该方法在煤矿安全监测中具有潜在的应用价值； 其次提出应用时间趋势结构序列数据挖掘方法，进行煤矿井下瓦斯时间序列趋势的 预测，MATLAB 仿真实验结果表明，时间序列数据挖掘方法在煤矿井下环境预测中 具有重要的应用价值；最后提出应用特征分析方法进行井下监测传感器的优化 布 点，建立了特征分析模型，突破了传统传感器布点只依据《煤矿安全规程》的布点 方法，在理论和科学的层次上对传感器的布点问题进行了研究，实现了井下环境监 测传感器的灵活布置。MATLAB 仿真实验验证了方法的可行性。但由于条件和时间 的限制，很多方面还有待改进和深入研究。 关键词：数据挖掘；安全监测；模糊聚类；模糊评判；时间序列预测；特征分析。 Abstract The application of the technology of data mining in the mine safety monitor system is investigated and discussed. Aimed at the deficiency of our country’s coal safety mining system, complex CAN bus and RFID technology are proposed and used in coal mine safety monitor system and achieve some of the hardware design. This method can’t only compensate for the traditional cable transmission measurement of gas concentration in a fixed place of issue of the lack of mobility, but also solve the wireless sensor networks with limited power supply of the state. It is Combines the merits of both approaches can be better monitored. On the software side: First proposed the application of Data Mining fuzzy clustering method and fuzzy evaluation technology for underground division of the regional environmental monitoring and risk levels determined, a new statistic are proposed complex distance measure and similarity measure.MATLAB simulation results show that the method in the coal mine safety monitoring has a potential application value; Second, application of TSDM method to predicte the gas trend in the future. MATLAB simulation results show that time series Data Ming method for eniviromental prediction in the coal mine has an important application value; Last, application of Characteristic analysis method to optimize the place point of the sensor, breaking the traditional distribu