Vol.9 No.2 2016
37/62

Research paper : Constructing a system to explore shallow velocity structures using a miniature microtremor array (I. CHO et al.)−93−Synthesiology - English edition Vol.9 No.2 (2016) 4.1.3 Three-point irregular array analysis (measure: evaluation of irregular array layout)In cases where the analysis results of a miniature array with a radius of 0.6 m are of low quality, arrays with a larger size are used to supplement the data. Here, arrays larger by one order (radii of several meters) are considered. However, arrays of this size do not fit behind a car parked on the roadside. Conventionally, arrays of Fig. 1(b) where the seismometers are laid out at equal intervals on a circle are used, but due to the problem of roads and buildings in the urban areas, one cannot expect the arrays to be laid out neatly.As a measure for this, we considered using three-point irregular arrays [Fig. 1(c)] of a radius of about 5 m as supplementary arrays. In conventional microtremor array exploration, data processing is done assuming that arrays are laid out at equal intervals, and it is difficult to apply such three-point irregular arrays. However, the authors have already proposed a data processing method in cases of unequal intervals, and have theoretically evaluated the effect on the analysis results.[13] In this research, using an actual equipment,[2] field experiments were repeatedly conducted by combining miniature arrays and three-point irregular arrays of several meters size, and the applicability to the new system was studied. As a result, it was shown that the supplementary data that fulfill the objective could be obtained by applying this idea, and that it is sufficiently practical without much difference in measurement time and amount of labor.[22]4.1.4 Estimation of velocity structure (measure: use of simplified analysis)What we ultimately want is the S-wave velocity structure of the ground. To model the velocity structure, nearby logging data and wide-area geological structures are referenced as initial models, and in many cases, iterative inversion analyses are conducted to optimize a model to fit the phase velocity obtained in array measurement.However, expert knowledge is necessary to create the initial models, and this does not match the demands for automatic processing and instant delivery of analysis results. Therefore, here, a simple conversion method is adopted, where the dispersion curve that is expressed as frequency versus phase velocity is converted to the relationship between wavelength and phase velocity, and this is deemed as the relationship between depth and S-wave velocity through appropriate scaling.[22] A method where the phase velocity that corresponds to the wavelength of 40 m is deemed as the average S-wave velocity from the surface to a depth of 30 m is also used concurrently.[23]This falls within the range of a standard preliminary analysis, but we believe that analysis results in a sufficiently satisfying range can be obtained (for example, Subchapter 4.3) when considering the balance between supply and demand. However, currently, we are also investigating the possibility of some advanced inversion methods such as a simultaneous inversion[24] of the H/V spectrum and the dispersion curve.4.1.5 Automation of the data analysis (measure: optimization of the parameter)From the perspective of use by general users, all processes including the selection of analysis intervals[19] and reading of phase velocity (Fig. 6)[25][26] should be automated. This doubles the measures for maintaining objectivity in reading, time reduction, and human error countering. The design also allows visual checking of the results of automated processing afterwards, and re-reading.In the estimation of velocity structures, the analysis results for each measurement point are presented, but in many cases the interpretation is shaky with only those results. It will be very effective in supplementing the interpretation if there is a two-dimensional velocity cross-section made by connecting together the results of multiple measurement points. Therefore, we created a tool that automatically draws the S-wave velocity cross-section simply by setting the cross-section line on the map (Subchapter 4.3).4.1.6 Quality control of the raw data and analysis results (measure: creation of index)In cases where there are problems in the raw data due to lack of data, failure to synchronize time on GPS, poor installation environment, or others, the decision of whether to conduct re-measurement can be made on the spot, by presenting these facts in quasi-real-time. In cases where the quality of the analysis results is low (such as when exploration does not reach the target depth due to low SN ratio since the microtremor intensity is low, etc.), deciding to conduct additional measurement by three-point irregular arrays becomes possible. To avoid overinterpretation of the analysis results, it is necessary to conduct thorough quality control. The index for quality control will be presented by combining the maximum amplitude of waveform, spectral density, SN ratio, and other quantities. Details are currently being investigated.4.2 Development of the measurement system4.2.1 Development of the seismometer (measure: maintaining mobility, robustness, stability, and simplicity)The seismometer used in the microtremor array exploration must have small individual differences in response characteristics, and they must be particularly uniform in phase characteristics. Also, to obtain high quality data in regions with small microtremor levels, it is important to have low self-noise. Considering the possibility that they may be used for microtremor arrays of several hundred meters size, a wide-frequency-range seismometer that has

元のページ 

page 37

※このページを正しく表示するにはFlashPlayer10.2以上が必要です