影响矿井通风实时仿真的风流随机扰动识别研究

The speed and precision of mine ventilation network solution have been able to satisfy the requirements of mine ventilation engineering. When the solution is applied in coal mine, there is still great error between the solution and real-time monitoring data of mine monitoring and supervision system, also ventilation network simulation system can not be connected with monitoring and supervision system to carry out real-time simulation. This situation is due to air flow fluctuating all the time, but turbulence fluctuation and random disturbance caused by mine production activity are the main reasons for the variation of air flow fluctuation. Experimental studies have showed that, air flow velocity of the disturbance source tunnel and neighbors will fluctuate in an instantaneous or short period of time, when the ventilation system is disturbed by the production activities such as the air door switch, tramcar moving, cage hoisting, etc. If we can identify air flow velocity variation caused by the disturbance, so the monitoring data of air flow velocity can be corrected, thus keeping the consistency to ventilation network solution. In order to study the regularity of air flow velocity variation, an experimental model would be designed to simulate typical mine ventilation system, coming with attachments of various random disturbance sources such as air door switch, mine car moving, cage hoisting, etc. Also some professional precise instrument are equipped to monitor air flow velocity, such as Laser Doppler Anemometer, Hot Wire Anemometer, etc. The implicit characteristics and recognition method of time-averaged air flow velocity disturbance wave would be studied after the turbulence fluctuation was filtered out, inspired by power quality electrical signal disturbance recognition.

矿井通风网络解算方法的速度和精度已经能够满足矿山通风工程要求，但在矿山现场应用时，实际解算结果与矿井监测监控系统的实时监测数据仍有较大误差，亦即网络解算仿真系统尚不能与监测监控系统联网进行实时仿真。其原因在于井下风流每时每刻都在波动变化，而湍流脉动和矿井生产活动产生的随机扰动是风流波动变化的主要原因。课题实验研究发现，当通风系统受到风门开关、矿车运行、罐笼提升等生产活动扰动时，扰源巷道及相邻关联巷道的风速会发生瞬时或者短时周期性的波动变化。如果能够识别这种扰动产生的风速变化，就可以对风速监测数据进行修正，从而保持监测数据与网络解算数据的一致性。课题在实验室搭建典型矿井通风系统模型，并设计风门开关、矿车运行等扰动源，通过LDA、HWA等精密测试仪器，监测巷道风速变化规律。借鉴电能质量电信号扰动识别的思想，研究过滤湍流脉动后的时均风速扰波隐含特征及其识别方法。

当通风系统受到风门开关、矿车运行、罐笼提升等生产活动扰动时，扰源巷道及相邻关联巷道的风速会发生波动变化。如果能够识别这种扰动产生的风速变化，就可以对风速监测数据进行修正，从而保持监测数据与网络解算数据的一致性。项目综合采用实验研究、理论分析与数值模拟相结合的方法，探析了风门开闭、矿车运行扰源主要参变量对扰源所在巷道以及其它关联巷道的风流扰动影响规律；根据实验数据建立了基于机器学习理论的风门开闭单扰源扰动分类识别模型，提出了基于统计分析和离散小波变换的扰动特征提取方法，以及基于多尺度滑动窗口和支持向量机的风门开闭单扰源扰动识别方法。该工作对提高矿井风速在线监测精度具有重要的理论意义和实用价值。