【作者】 巩思园；

【导师】 窦林名；

【作者基本信息】 中国矿业大学 ， 采矿工程， 2010， 博士

【摘要】 针对进入深部开采冲击矿压日趋严重的现实,对获得深部煤岩体的应力信息,实现冲击矿压的预测预报显得更加迫切。论文围绕实验室研究得到的震动波波速与应力间的试验关系模型,利用矿震信号建立了矿震震动波速层析成像模型,在分析该模型各影响因素的基础上,采用理论分析、现场实验、数值模拟和数值仿真计算等方法对台网优化布设、提高输入数据精度进行了研究,进而揭示了冲击矿压与震动波波速的耦合规律,提出了预测预报冲击危险的三个指标值和防治对策,并在现场进行了工程实践应用。实验室研究了不同加卸载方式下应力与波速的耦合关系,建立了两者之间的试验关系模型,并对模型参数进行了分析。研究表明:三轴和单轴条件下,煤岩试样在弹性阶段纵波波速变化梯度大,塑性阶段波速变化趋于平缓,应力与纵波波速间具有幂函数关系,与实测值的相关系数计算说明模型相对理论公式具有更高的拟合度。基于矿震震动波速层析成像理论,建立了阻尼最小二乘求解模型,提出采用SIRT算法和多种射线追踪方法求解震动波波速。为评价反演结果稳定性,创建了基于统计学方法的波速偏离期望值评价模型。研究表明,模型虽求解速度慢,但能够稳定收敛,评价模型与射线覆盖密度存在很好对应关系。应用D值优化理论研究了影响矿震定位精度的主要因素,提出采用综合指数法确定危险区域和矿震发生概率的方法,通过理论分析震中和震源标准差反映台网定位能力的不足,建立了震中与震源误差期望值模型,最终形成台网布置优化及评价系统。实验和现场应用结果表明,该系统能够快速确定台网最优布置方案,准确评价台网定位能力。为提高矿震震动波速层析成像输入数据精度,确定利用纵波到时定位,并采用小波进行滤波提高到时标记精度,以此,创建GE-POWELL算法,确定求解的最优通道个数和初始纵波波速,建立反演计算中的两种震源求解模型。现场实验结果表明以上方法可有效提高震源的定位精度。基于强度、能量理论及动载荷关系式,论文创建了波速异常、波速梯度变化VG值和由试验关系模型确定的应力集中系数三个指标预测预报冲击危险,提出了防治措施,并在济宁三号井163下02C孤岛工作面进行了CT实验与数值模拟的冲击危险对比分析。对比结果表明:震动波速与数值模拟存在不同点和相同点,并比数值模拟的适用性更强。基于SOS监测系统,利用忻州窑煤矿8929工作面和鲍店煤矿103上02工作面回采产生的矿震信号进行冲击危险预测预报和防治实践,并采用三个冲击危险指标预测预报了下一个时段的冲击危险区域及危险级别,未来监测结果显示冲击矿压或强矿震大部分都发生在预测的危险区域内,并且很好的对应了煤柱应力集中区等位置,表明矿震震动波速层析成像用于冲击危险预测预报是可行且准确的,并为现场冲击矿压防治指明方向。更多还原

【Abstract】 The rockburst phenomena become progressively more severe as the mining depth and extent increase, so it is urgent to get the stress information to realize the prediction of rockburst. Focusing the experimental relationship between stress and tremor velocity derived from laboratory investigation, this thesis built the mine tremor velocity tomography model to calculate the velocity using mine tremor information. The influencing factors of the model,such as the optimal configuration of seismological observation network and the improvement of input data precision, were analyzed by theoretical analysis, field testing, numerical simulation and numerical emulation so as to obtain accurate result, thereby revealing the correlation between rockburst and velocity, which is used to guide the prevention measures and calculate three prediction indexs. At last, the method was applied to the engineering practice and applications.Through the laboratorial research, the experimental models between stress and tremor velocity were created under different loading and unloading ways. Also the parameters of the derived model were analyzed. Studies indicate that the velocity gradient is usually highest at elastic stage and then begins to level out at plastic stage. Velocity can be used to infer stress distribution because the relationship can be expressed as power functions. The measured sonic velocity showed that the experimental models have the better fitness than theoretical formula.The damping least square solution model is proposed based on the mine tremor velocity tomography theory, wich use SIRT algorithm and various seismic raytracing methods to calculate the velocity. In order to evaluate the stability of inversion result, the evaluation model of expected velocity deviation was created based on the statistical methods. Studies show that the solution algorithm is rather time-cousuming, but can ensure the program to be convergent steadily, and also the good corresponding relationship exists between the expected velocity deviation model and raypath sampling density.In this paper, according to practical conditions and requirements of coal mines, especially the accuracy of picking seismic wave arrivals at stations and building velocity model, the main influencing factors and bad conditions which influence the accuracy of source location were analyzed using the theory of D-optimal design and localization. A view of using the comprehensive index method to determine the high microseismic activity zones and the occurrence probability of mining-induced tremors was also proposed. In addition, since the location capacity of network can’t be reflected accurately using the source standard error calculated by the traditional theoretical analysis, the model of expected epicenter and hypocenter error was created by numerical emulation experiment, and then the optimal design and assessment system of network was built up finally. The results from experiment and field practice show that, the optimal design and assessment system can be used to solve the optimal plan of network quickly and evaluate the location ability of network exactly.In order to improve the input data precision, the P-wave was selected and filted by wavelet analysis to improve the accuracy of picking first arrival time which is used in the two kinds of source location algorithm in forward calculation by the created algorithm GE-POWELL to identify the best channel number and the initial velocity value. The application of these methods in field tests can effectively improve the location accuracy of mine tremor.Based on strength theory, energy theory and dynamic load equation, three prediction indexs including velocity anomaly, velocity gradient VG anomaly and stress concentration factor calculated by experimental relationship between stress and tremor velocity were created to predict rockburst danger. The prediction results are used to guide the prevention measures. The compared analysis between CT experiment and numerical simulation was conducted in LW16302C Jisan Coal Mine. The application results show that they have the common points and differences, but velocity has the stronger applicability.The on-site prediction and prevention practice was conducted by using the recorded mine tremor signals form LW10302 in Baodian Coal Mine and LW8929 in Xin Zhouyao Coal Mine with SOS microseismic monitoring system. The three prediction indexs calculated from inversion velocity distribution were used to predict the rockburst danger zone and level for the next period and the future monitoring results showed that most of the rockburst events and strong mine tremors truly occurred in the predicted danger zones and well correlated with the stress concentration zones in coal pillar etc. The successful application indicated that the mine tremor velocity tomography can exactly predicit the rockburst danger and point out the direction for prevention.更多还原