Study on simple estimate method of soil parameters and applicability of simplified equation of attenuation with distance in prediction of ground vibration

本研究围绕以下五个方面展开：1）简易激振装置的制作; 2）估计系统的建立; 3）估计系统的改进; 4）衰减随距离简化方程的适用性; 5）卓越频率的测量。总质量约为30公斤的激振设备是由落重和弹簧组合而成的。在第2项中，将地基模拟为粘弹性半空间，利用第1项中得到的激振力波形，建立了估计土壤参数的神经网络系统。第三项是对神经网络的改进。研究了用水平位移代替时间滞后是否改进了估计系统。从几个土壤剖面的结果来看，发现从输入参数中排除时间滞后并不会导致系统的改进。在第4项中，检验了Bronitz方程中的系数与第3项中估计的土壤参数的相关性。结果表明，由估计的土壤参数得到的系数的Bornitz方程可用于预测地面振动随距离的衰减，第5项是研究是否可能从白天测得的地面振动中提取出场地的卓越频率，并利用卓越频率，从简单预测方法的观点出发，提出了预测地面振动的频率。

This research is performed on the following five items ; 1)the manufacture of simple exciting equipment, 2)the establishment of estimate system, 3)the improvement of estimate system, 4)the applicability of simplified equation of attenuation with distance, 5)the measurement of predominant frequency.In the item 1, the exciting equipment with the total mass of about 30 kg is manufactured by combining a falling weight and a spring. It is examined through the field measurement that the equipment generates exciting force about equal to expected force.In the item 2, a neural network system is built to estimate soil parameters in which the ground is modeled as a visco-elastic half-space and the waveform of exciting force obtained in the item I is used. It is confirmed that this estimate system is possibly applied to practical use.In the item 3, the improvement of neural network is treated.. It is examined whether the estimate system is improved by using horizontal displacements instead of time lags. From results for several soil profiles, it is found that it doesn't lead to improvement of the system to exclude the time lags from input parameters.In the item 4, it is examined to relate coefficients in Bronitz equation to soil parameters estimated in the item 3. It is found that Bornitz equation with the coefficients obtained from estimated soil parameters may be applicable to prediction of the attenuation with distance of ground vibration.In the item 5, it is performed to find out whether it is possible to pick out predominant frequency at a site from ground vibrations measured in the daytime and to use the predominant, frequency for the prediction of ground vibration from the viewpoint of simple prediction method.