Earthquake Reaearch in China  2019, Vol. 33 Issue (2): 354-366     DOI: 10.19743/j.cnki.0891-4176.201902003
Characteristics of the Seismic Waves from a New Active Source Based on Methane Gaseous Detonation
WANG Weitao1, WANG Xiang2,3, MENG Chuanmin2,3, DONG Shi2,3, WANG Zhigang2,3, XIE Junju1, WANG Baoshan5, YANG Wei1, XU Shanhui1, WANG Tao4
1. Key Laboratory of Seismic Observation and Geophysical Imaging, Institute of Geophysics, CEA, Beijing 100081, China;
2. National Key Laboratory of Shock Wave and Detonation Physics, Institute of Fluid Physics, CAEP, Mianyang 621999, Sichuan, China;
3. Joint Laboratory of High Pressure Physics and Seismology, Institute of Fluid Physics, CAEP, Mianyang 621999, Sichuan, China;
4. Institute of Geophysics and Geodynamics, School of Earth Sciences and Engineering, Nanjing University, Nanjing 210046, China;
5. School of Earth and Space Sciences, University of Science and Technology of China, Hefei 230026, China
Abstract: Active seismic sources are critical for obtaining high resolution images of the subsurface. For active imaging in urban areas, environment friendly and green seismic sources are required. In present work, we introduce a new type of green active source based on the gaseous detonation of methane and oxygen. When fired in a closed container, the chemical reaction, i.e. gaseous detonation, will produce high pressure air over 150MPa. Seismic waves are produced when high pressure air is quickly released to impact the surroundings. The first field experiment of this active source was carried out in December, 2017 in Jingdezhen, Jiangxi Province, where a series of active sources were excited to explore their potential in mine exploration. In current work, we analyzed the seismic waves recorded by near-field accelerators and a dense short-period seismic array and compared them with those from a mobile airgun source, another kind of active source by releasing high pressure air into water. The results demonstrate that it can be used for high resolution near surface imaging. Firstly, the gaseous detonation productions are harmless CO2 and water, making it a green explosive source. Secondly, the dominant seismic frequencies are 10-80Hz and a single shot can be recorded up to 15km, making it suitable for local structure investigations. Thirdly, it can be excited in vertical wells, similar to traditional powder explosive sources. It can also act as an additional on-land active source to airgun sources, which requires a suitable water body as intermediate media to generate repeating signals. Moreover, the short duration and high frequency signature of the source signals make it safe with no damage to nearby buildings. These make it convenient to excite in urban areas. As a new explosive source, the excitation equipment and conditions, such as gas ratio, sink depth and air-releasing directions, need further investigation to improve seismic wave generation efficiency.
Key words: Active source     Seismic waves     Methane detonation source     Green seismic source

INTRODUCTION

Seismic wave is an important tool for studying the internal structure of the Earth, as bright lights that illuminate the subsurface (Chen Yong et al., 2005). The seismic waves released by natural earthquakes are very energetic and capable of traveling tens of thousands of kilometers, making it suitable for studying large-scale structures of the Earth's interior. The active source, compared to natural earthquakes, has the advantages of controllable time and position, thus is widely used in the field of crust structure imaging and mineral resource exploration (Mooney W. D. et al., 2002).

When the observing system is sufficiently dense, the characteristics of the seismic wave from the active source, such as its propagation distance and dominant frequency, determine the range and resolution of the crust structure imaging. In terms of P-wave signals commonly used in active source exploration, seismic waves of higher frequencies have smaller wavelengths, which help to obtain fine images of underground structures. However, high-frequency seismic waves attenuate faster when propagating in the medium. In order to obtain a clear record at a longer propagation distance, it is necessary to increase the released energy of the active sources. Sources with higher single-excitation energy tend to cause damage near the excitation point, which restricts its usage where none-destructive exploration is required. The selection and use of active sources requires a balance between propagation distance and imaging resolution according to the research objectives.

The explosive source is the most widely used active source. Deep seismic sounding techniques use large equivalent explosive sources to obtain velocity structures of the crust and topmost upper mantle (Zhang Zhongjie et al., 2011; Dong Shuwen et al., 2013). In mineral exploration, small equivalent explosive sources are densely excited to obtain fine local structures. However, as explosives will release harmful chemicals and pollute the environment, they are increasingly restricted. As a result, researchers have developed seismic sources based on continuous vibrations such as Minisose, Vibroseis and ACROSS, to replace chemical blasting to generate seismic waves (Yamaoka K. et al., 2001; Chang Xu et al., 2008; Wang Hongti et al., 2009). Compared with the explosive source, these sources have lower energy conversion efficiency on exciting seismic waves. Wang Baoshan et al. (2016) used a large-volume airgun array to generate seismic waves, which releases high-pressure gas in the in-land water body suddenly to generate high repeating signals. The repeatability is explored to increase the equivalent excitation energy by stacking many single shots. Large-volume airgun can excite seismic waves with a dominant frequency of about 5-8Hz and can propagate 1, 000km after a thousand stacking (Chen Yong et al., 2017). To obtain high repeatability, the airgun source excitation requires the water body as a medium, and its use will be limited when a suitable water body is not available.

Oxygen and combustible gases (such as hydrogen, methane, etc.) can be mixed and will undergo a detonation reaction under certain energy ignition conditions. The gas mixture can generate high temperatures and high pressure gas through a detonation reaction in a closed vessel. Similar to the airgun source, a rapid release of the high pressure gas can form a shock wave. This technology is currently widely used in high-speed propulsion technology research (Zhang Bo et al., 2014; Dong Shi et al., 2017). Taking the mixture of methane and oxygen as an example, when a detonation reaction occurs, a large amount of chemical energy is released and the products are water and carbon dioxide, which are favorable for environmental protection. Therefore, methane and oxygen can be used to produce a gas phase detonation in a specific device and release high pressure gas to impact the medium as a new type of seismic source. The excitation of this source is similar to the explosive source and it can be excited in the onshore well, reducing the dependence of the water body.

In December, 2017, several institutions carried out an experiment in Jingdezhen, Jiangxi, to explore the ore-forming structure of the area using active sources and dense observation systems. The methane and oxygen mixed gas detonation source (hereinafter referred to as the methane detonation source) was first used during the experiment. In this paper, the seismic waves excited by the methane detonation source are analyzed, and the characteristics of near-field strong motion, dominant frequency and propagation distance are obtained. These results show the main characters of this new source.

1 METHANE DETONATION SOURCE EXCITATION AND OBSERVATION EXPERIMENTS

The Zhuxi area of Jingdezhen City in Jiangxi Province is located in the eastern part of the Jiangnan orogenic belt. In recent years, the tungsten ore body associated with copper and zinc has been discovered, which has changed the mineral distribution pattern of "South Tungsten North Copper" in Jiangxi and its neighboring areas (Chen Guohua et al., 2012). In December 2017, researchers used a variety of active sources to conduct active exploration experiments in the Zhuxi mining area to study the geological background of their mineralization. A dense array composed of 178 short-period three component seismometers in the 30×40km area of the Zhuxi mining area was deployed with a station spacing of 1-2km (Fig. 1(a)). The continuous observation lasted for 30 days at a sampling rate of 200Hz. At the same time, four dense reflection lines with an averaged spacing of 10m were laid in the vicinity of the mining wells with nearly 2, 000 receivers. A variety of active sources, such as vibrators, small airgun sources and hammers, were used to excite seismic waves to feed the observation system.

 Fig. 1 Map of the sources and receivers in the Jiangxi active source experiment (a) Map of the stations. In total, 178 short-period seismometers (black solid triangles) and four dense receiver lines (four color lines) are deployed. The red rectangle indicates region shown in (b)). (b) Source location of the mobile airgun (black pentagram) and methane detonation source (red pentagram). The four dense receiver lines are marked from L1 to L4. The locations of methane detonation source and three accelerators are shown in the insert

The dense and diverse recording system in the Zhuxi area is favorable to study the seismic waves from the new methane detonation source. During the experiment, the methane detonation source was excited at the intersection of the lines L2 and L3 (Fig. 1(b) red pentagram). At the same time, three CMG-5TDE accelerometers were installed at 6m, 30m and 280m away from the source point to record the near-field strong motions at a sampling rate of 200Hz. Using these observation systems, the seismic waves excited by the methane detonation source can be well recorded and analyzed.

Fig. 2(a) shows the device for the methane detonation source. The device is a cylindrical steel sealed container, one end of which has an injection valve and an ignition initiation interface. When the source is excited, through the injection slot, methane and oxygen are injected and mixed at the reaction ratio in the sealed container, then a detonation reaction occurs after ignition to generate a high-pressure gas in the container. A pressure limiting valve is arranged at the other end of the container, the internal high pressure gas is released when the pressure exceeds the threshold. Similar to the traditional explosive source, the source needs to be placed in the well when it is excited, as shown in Fig. 2(b). The quick release of the high-pressure gas can impact the surroundings to generate seismic waves. In this experiment, the methane detonation source device has an outer diameter of 120mm and a length of 1000mm. The unit was placed in a well with an outer diameter of 120mm and the bottom of the unit was 6m from the wellhead. The initial gas injection pressure of methane and oxygen is 6MPa, and the internal pressure can reach 165MPa after ignition. The pressure limiting valve is designed to be located at the bottom of the vessel to ensure that the main impact energy travels downward.

 Fig. 2 Instruments and elastic wave generation mechanism of the methane detonation source (a) Image of the methane detonation source equipment. (b) Schematic illustration on the deployment and excitation of methane active source
2 CHARACTERISTICS OF SEISMIC WAVES EXCITED BY METHANE DETONATION SOURCE

The near-filed strong motion strength of the active source affects its choice of excitation environment. The propagation distance and dominant frequency of the excited seismic waves control the range and resolution of seismic imaging. We use the records from near-field accelerators, short-period array and dense receiver lines to analyze the seismic signatures of the methane source.

The released energy, well sealing condition and coupling effect of the source affects near field ground acceleration. Fig. 3(a) shows the acceleration waveform and their frequency characteristics at 6m, 30m and 280m from the methane detonation source excitation point. The near-field vibration has a short duration of about 0.25s at 6m and about 0.3s at 30m. At longer distances, the signal is coupled to the subsurface structure, and its waveform becomes complex, lasting about 0.5s at 280m. The spectrum of the vibration is wide, with a corresponding energy distribution at 10-80Hz and a peak at 20-60Hz. The intensity of the high-frequency vibration signal decays rapidly with distance. At 6m from the source, the peak acceleration reaches 648, 270 and 1090Gal in the east-west (E), north-south (N), and vertical (Z) directions, respectively. At 30m, the peak acceleration decays rapidly to 84, 38, and 79Gal. It decays below 10Gal at 280m, more specifically, to 9.1, 5.1 and 6.2Gal, respectively. Due to the fact that the well is not well sealed during the excitation, the source has significant recoil movement during the excitation, which may cause the higher acceleration of the Z component at 6m. At the same time, the signal strength at 280m is attenuated to about 1/8 compared to the signal at 30m, but its high-frequency signal ratio is increased, which may be caused by the coupling with the complex shallow structure during the propagation.

 Fig. 3 Characteristics of nearfield strong motions generated by the methane detonation source (a) The waveforms for ground acceleration at distances of 6m, 30m and 280m. (b) The attenuation curves of peak ground accelerations (PGA). The east-west, south-north and vertical components are indicated as red, blue and black respectively

The earthquake ground motion response spectrum can reflect the influence of strong ground motion on building structures with different self-vibration periods. We compare the near-field horizontal acceleration recording response spectrum of the methane detonation source with the near-fault records obtained from the Wenchuan and Lushan earthquakes to analyze the near-field effects of the methane detonation source. Fig. 4 shows the response spectrum curves of the horizontal E component acceleration recording at 6m and 30m of the gas detonation source event. For comparison, the curves at Wolong station (WCW) of the Wenchuan MW7.9 earthquake and at Baoxing station (BXD) records of the Lushan MW6.8 earthquake as well as those for seismic design of buildings (China's Codes 50011-2010) are presented. Compared with the natural earthquake response spectrum, although the peak of the response spectrum of methane detonation source at 6m is as high as ~2.5g, which is close to the near-fault record in the Wenchuan MW7.9 and Lushan MW6.8 earthquakes, the peak period is only 0.03s. The response spectrum value decays rapidly with the increase of the period, and is much smaller than the actual record of natural earthquakes in the building structure sensitive 0.1-2.0s period, and lower than the design spectrum level of the Ⅸ degree area and the Ⅷ degree area required in the seismic design specification. The high frequency signal from the methane detonation source decays very quickly. The maximum response spectrum is less than 0.2g at 30m and the peak period is 0.03s. At 280m, it is less than 0.01g, thus not shown in the figure. The short duration and high frequency characteristics make the response spectrum value much smaller than the seismic design spectrum level in the building structure sensitive 0.1-2.0s period. The methane source tends to be an environmentally friendly source with little damage to the near-field buildings.

 Fig. 4 Comparison of observed response spectra generated from methane gaseous detonation with the observation in Lushan (red thin line) and Wenchuan (green dashed line) earthquake and design spectra of the code for seismic design of buildings (50011-2010) The maximum considered horizontal spectral accelerations in the design spectra are 1.4g and 0.9g for seismic intensity IX (gray solid line) and Ⅷ (gray dashed line), respectively. The horizontal period axis is shown in log scale

The seismic waves excited by the methane detonation source are well recorded by the short-period array and dense receiver lines. Fig. 5 shows the vertical components records of the dense array and the L3 line after 10-80Hz bandpass filtering. As the layout of the dense array is dense inside and sparse outside (Fig. 1(a)), the records of methane detonation source is dense in the range of 13km from the epicenter, and is sparse outside. Fig. 5(a) shows that, in the 13km range, clear P-wave signals are tractable in the short-period records. The P-wave signal was observed at a station about 15.5km away from the epicenter, but at 18km, the P-wave signal strength was only slightly higher than the noise. Therefore, the methane detonation source can achieve a horizontal propagation distance for at least 15km. The P-wave signal can also be continuously tracked on the receiver line with a longest offset of 6.8km. The P-wave signal generated by the methane detonation source is strong with an averaging speed about 5.1km/s. Compared with the P-wave signal, the S-wave signal is weaker. This shows that, similar to the explosion source, the seismic waves by the vertical downward high-pressure gas shock is mainly P-wave signals, while the shear component is weak.

 Fig. 5 Records of the seismic waves generated by the methane detonation source (a) Seismic waves from methane detonation source recorded by the dense short-period array. (b) Seismic waves from methane detonation source recorded by dense receiver line L3. All the waveforms are vertical component and 10-80Hz band-pass filtered

The airgun source is also an active source that uses the sudden release of the high-pressure gas to generate seismic waves. However, it uses the water body as an intermediate medium, and the characteristics of excited seismic waves are different from those from the methane detonation source. During the experiment, we used a small airgun with a capacity of 250 cubic inches (0.0040968 cubic meters) to perform 215 excitations in the Shantianwu Reservoir (black pentagram in Fig. 1(b)). The small airgun was placed at 4m under water during excitation. The excitation pressure is 15MPa. The experiment was divided into two stages, using two different airgun control devices with a slightly difference on the airgun valve closing speed. In the first stage, the airgun was excited 176 times in 9 hours; and in the second stage, it was excited 39 times in 3 hours (Fig. 6).

 Fig. 6 The repeatability of the signals generated by mobile airgun source Top panel: The cross-correlation coefficients between the linearly stacked reference signal and that from each single shot. The coefficients for 0.96, 0.98 and 0.99 are represented by the dotted, dashed and solid horizontal lines, respectively; Bottom Panel: The linearly stacked references waveforms are shown as thick lines and those from single shots are in thin lines at two different stages. The station is about 200 meters away from the airgun source. All waveforms are band pass filtered between 8-15Hz. See text for details

The repeatability of the signal generated by the airgun source is an important feature that distinguishes it from other active sources. We selected a station 200m away from the airgun source excitation point to analyze the airgun source repeatability. The bottom panel of Fig. 6 shows the vertical records of 154 single shots in the first stage and 39 single shots in the second stage (not all 215 excitation signals are recorded by this station due to significant human interference). The single shot records of the two stages were linearly stacked as reference waveforms, respectively, as shown in the bold waveforms in Fig. 6. The cross-correlation coefficients between the reference waveform and the corresponding single shot waveform can represent the source repeatability. The window for cross-correlation calculation is selected as 0-0.4s, and the data is band pass filtered between 8-15Hz. The resulting cross-correlation coefficients are shown in Fig. 6. It can be seen from Fig. 6 that the cross-correlation coefficients between the reference waveform and the single shots in both stages are above 0.96, which shows the high repeatability of the airgun excited signal. At the same time, the coefficients gradually increase with the number of excitations, eventually exceeding 0.99. This is due to the fact that the water body for excitation is a small shallow reservoir. In the initial stages, the airgun source that is placed at a depth of 4m will cause slight changes in the excitation environment such as sludge at the bottom of the reservoir. As the number of excitations increases, the excitation environment tends to be stable, and the repeatability of the waveform is further increased. The cross-correlation coefficient of the two reference waveforms is 0.94, indicating that the two control devices have little influence on the repeatability. Therefore, in the subsequent stacking process, the shots of the two stages are processed together.

Fig. 7 shows the vertical waveform and the time-frequency distribution of the airgun signal at the nearest seismometer 200m away from the airgun source. Unlike the methane detonation source, the energy of the small airgun source is mainly below 30 Hz, where the high frequency 15-28Hz is the coupling of the high pressure gas to the bottom of the reservoir after impacting the water body, and the 8-15Hz is generated from the oscillation of the bubble in the water body. Because of the high repeatability, we stack the 215 airgun shots recorded by the dense array, and the 8-15Hz filtered waveform is shown as Fig. 6(c). Thanks to the lower frequency and the random noise suppression through stacking, the stacked airgun signal are observed by all stations up to 20km. Compared with the methane detonation source, the airgun source has a lower signal frequency and a longer propagation distance. At the same time, compared with the methane detonation source, the S-wave signal in the airgun signal is stronger, which may be caused by the conversion of seismic waves on the bottom and boundary of the reservoir (Chen Meng et al., 2013; Hu Jiupeng et al., 2017; Sun Nan et al., 2017).

 Fig. 7 The time-frequency representation of the airgun generated seismic signal at nearfield and records of the propagating seismic waves by the dense array (a) The vertical records of the airgun signals at the nearest seismometer 200m away from the airgun source. (b) The time frequency representation of the waveform in a), the energy is normalized to 1.0. (c) The vertical records from short period seismic array for stacked airgun signals. In total, 215 airgun shots are linearly stacked then band-pass filtered between 8-15Hz

In the seismic waves excited by the methane detonation source and the airgun source, the dominant frequency and relative strength of various phases are different. We compare the records from station 21711 which has similar distances from the two sources. The station is 5.856km away from the methane detonation source and 5.403km away from the airgun source. Fig. 8(a), (c) shows the vertical component records at station 21711 for the methane detonation source signal and the stacked airgun signals from 215 shots, respectively. The waveforms are band-pass filtered using a broad band 1-80Hz to retain more information. To compensate for the difference in epicentral distance, the waveforms are aligned using corresponding P-wave arrivals (as mark time zeros) for easy comparison. Fig. 8(b), (d) shows the time-frequency representations of the corresponding signal, respectively. It can be seen from the time-frequency representations that the dominant frequency of the methane detonation source signal is higher, the P-wave energy is concentrated between 30-60Hz, while the airgun source signal is between 8-30Hz. The P-wave signal of the airgun source is divided into two frequency segments of 8-15Hz and 10-30Hz, respectively corresponding to bubble oscillation and direct impact on the bottom of the reservoir, which is consistent with the record at the near field of 200m (Fig. 7(b)). At a distance of about 5km from the epicenter, the relatively high-frequency signal formed by direct impact is not sufficiently attenuated, so its energy is higher than that generated by bubble oscillation. About 1 s after the arrival of the P-wave, the corresponding S-wave phase appears in both waveforms, but the amplitude ratio of the P-wave to S-wave of the methane detonation source is higher compared to the airgun source. About 0.6s after the P-wave, strong energy occurs in the waveforms and the time-frequency representations of both sources. The relevant phase may be caused by the reflection/refraction of the signal in the complex shallow structure, and its propagation mechanism remains to be studied. At the same time, the airgun source signal shows multiple energy bands above 40Hz, which may be caused by the accompanying signals generated by the complex shallow media and surface undulations.

 Fig. 8 The waveforms and time-frequency representations of the vertical records at the same station 21711, which has similar source-station distances from the mobile airgun and the methane detonation source (a) Vertical record for the methane detonation source generated signal.(b) Time-frequency representation of waveforms in (a).(c), (d) Same with (a) and (b) but for the linearly stacked records of 215 airgun shots.; The waveforms are band-pass filtered using a broad band 1-80Hz with corresponding distances labeled at the top-right corner of each panel. The waveforms are aligned using corresponding P-wave arrivals (as mark time zeros) for easy comparison
3 DISCUSSION AND CONCLUSION

Both the methane detonation source and the airgun source are active sources that generate vibration by releasing high-pressure gas. In order to obtain high-pressure gas, the airgun source uses physical means to compress atmospheric air to the gun chamber, while the methane detonation source uses chemical methods to obtain high-pressure gas by detonation reaction. Since the detonation reaction of methane and oxygen does not produce environmentally harmful substances, it is also a green seismic source, just like the airgun source. Both sources can provide environmentally friendly methods for the active imaging of underground structures.

From the perspective of energy conversion, the airgun source converts the potential energy of the high-pressure gas into impact energy, while the methane detonation source realizes the conversion of chemical energy. The energy of a single shot of airgun source can be calculated based on chamber capacity and pressure (Chen Yong et al., 2008):

 $E=P \cdot V \cdot \ln \left(P / P_{0}\right)$ (1)

Where E is the energy released by a single excitation, V is the volume of the airgun chamber, P is the pressure of the high-pressure gas, and P0 is the static pressure, which generally takes the standard atmospheric pressure of 0.1MPa. The calculation was carried out with a small airgun of 250 cubic inches (0.0040968m3) at a pressure of 15MPa. The energy released in a single excitation was 15 × 0.0041 × ln (15.0 / 0.1)=0.308MJ (terajoules). The energy density released by the dynamite is about 4kJ/g, and the energy released by a small airgun in single excitation is equivalent to 77g of dynamite.

The energy of the methane detonation source can be calculated from its initial pressure, gas ratio, and chamber capacity. In this experiment, the gas chamber volume was about 6 liters, with an initial pressure of 6.0MPa, and methane and oxygen were injected at the reaction equivalent ratio. According to the methane/oxygen reaction equation where g represents a gas phase reaction :

 $\mathrm{CH}_{4}(\mathrm{g})+2 \mathrm{O}_{2}(\mathrm{g})=\mathrm{CO}_{2}(\mathrm{g})+2 \mathrm{H}_{2} \mathrm{O}(\mathrm{g})$ (2)

Its reaction energy is ΔH=-850 kJ/mol. The total chemical energy of the detonation reaction was E=5.36mol × 850kJ/mol=4.56MJ. The chemical energy released by the methane and oxygen detonation reaction is not completely used for generating seismic waves, and a considerable part of the energy is converted into vibration and rotation of the product gas molecules. According to thermodynamic experiments and theoretical studies, the external functional capacity accounts for about 30% of the total chemical energy (Gao Rongqing, 1994). Therefore, the effective energy to generate seismic waves from the methane detonation source is about 1.4MJ in this experiment, equivalent to the energy released by 350g of dynamite. Due to the higher energy release density of chemical reaction, at the same volume and initial pressure, the chemical energy released by the methane detonation source is an order of magnitude higher than the potential energy released by the airgun source. In other words, the device for methane detonation source can be more compact when same desired energy is required.

The airgun source uses the water body as an intermediate medium to excite seismic waves, and has the characteristics of highly repeatable waveforms in multiple excitations. There is a high requirement of its excitation environment. The methane detonation source is similar to the traditional explosive source, and the impact is directly applied to the solid medium, which makes it easier to select the excitation environment. According to the characteristics of the excited seismic wave, the dominant frequency of the airgun source is lower, while the dominant frequency of the methane detonation source is higher. For this experiment, if the P-wave velocity is 5km/s and take 10Hz and 50Hz as the main frequency of small airgun and methane detonation source, the wavelengths are 500m and 100m, respectively. As a result, seismic waves excited by the methane detonation source have certain advantages in the resolution when detecting the underground structures. The airgun source can increase the detection depth by superimposing the detection distance by means of the repeatability of the waveform. Therefore, the two sources have certain complementarity in excitation conditions, detection range and precision. By exploiting the different characteristics of these two green seismic sources utilizing high-pressure gas, they can exert their respective advantages in the detection of shallow structures and form a series of gas phase active sources with complementary advantages.

As a new type of source, the excitation device and excitation technology of the methane detonation source still need further researches. By adjusting the initial pressure and ratio of the gas mixture, modifying the structure of the excitation device and the vent diameter, the energy and dominant frequency of the excited seismic wave can be adjusted. By adjusting the air outlet setting, the directional excitation of the gas impact can be achieved, thereby enhancing the seismic wave intensity in a specific direction.

In this paper, the characteristics of seismic waves excited by gas detonation sources are analyzed. The results show that the new source, based on methane and oxygen mixture, can generate high frequency seismic waves of 10-80Hz with horizontal propagation distances as far as 15km. At the same time, there is no damage to surrounding buildings, as the excited high-frequency seismic wave has a short duration and the near-field strong motions decay rapidly. These features make it a green seismic source for the active imaging of shallow crust structures, and can serve for non-destructive underground space exploration in urban and related areas.

REFERENCES
 Chang Xu, Li Linxin, Liu Yike, Wang Hongluo. Seismic profile of Huangzhuang-Gaoliying fault in Beijing by Mini-sosie method[J]. Chinese Journal of Geophysics, 2008, 51(5): 1503-1510 (in Chinese with English abstract). Chen Guohua, Wan Haozhang, Shu Liangshu, Zhang Cheng, Kang Chuan. An analysis on ore-controlling conditions and geological features of the Cu-W polymetallic ore deposit in the Zhuxi area of Jingdezhen, Jiangxi Province[J]. Acta Petrologica Sinica, 2012, 28(12): 3901-3914 (in Chinese with English abstract). Chen Meng, Wang Baoshan, Wang Weitao, Yang Wei, Song Lili, Lin Jianmin, Ge Hongkui, Chen Yong. A study on S wave and Poisson's ratio models in southern Yanshan uplift using seismic data generated by airguns in a water reservoir[J]. Progress in Geophysics, 2013, 28(1): 102-110 (in Chinese with English abstract). Chen Yong, Zhu Rixiang. Proposed project of "underground bright lump"[J]. Advance in Earth Science, 2005, 20(5): 485-489 (in Chinese with English abstract). Chen Yong, Liu Lanbo, Ge Hongkui, Liu Baojun, Qiu Xuelin. Using an airgun array in a land reservoir as the seismic source for seismotectonic studies in northern China:experiments and preliminary results[J]. Geophysical Prospecting, 2008, 56(4): 601-612. DOI:10.1111/j.1365-2478.2007.00679.x Chen Yong, Wang Baoshan, Yao Huajian. Seismic airgun exploration of continental crust structures[J]. Science China Earth Sciences, 2017, 60(10): 1739-1751. DOI:10.1007/s11430-016-9096-6 Dong Shi, Meng Chuanmin, Xiao Yuanlu, Mo Junjie, Zhang Mingjian, Wang Xiang, Shi Shangchun. Preliminary study of two-stage light gas gun using reactive gas as driving energy[J]. Chinese Journal of High Pressure Physics, 2017, 31(2): 182-186 (in Chinese with English abstract). Dong Shuwen, Li Tingdong, Lü Qingtian, Gao Rui, Yang Jingsui, Chen Xuanhua, Wei Wenbo, Zhou Qi, SinoProbe team. Progress in deep lithospheric exploration of the continental China:a review of the SinoProbe[J]. Tectonophysics, 2013, 606: 1-13. DOI:10.1016/j.tecto.2013.05.038 Gao Rongqing. The physical meaning of pyrotechnic force is analyzed from the kinetic theory of gas molecules[J]. Chinese Journal of Explosives & Propellants, 1994, 17(2): 40-44 (in Chinese). Hu Jiupeng, Wang Baoshan, Chen Yong. The influence of water shape on land airgun triggered signals[J]. Journal of Seismological Research, 2017, 40(4): 543-549 (in Chinese with English abstract). Mooney W.D., Prodehl C., Pavlenkova N.I. Seismic velocity structure of the continental lithosphere from controlled source data[J]. International Geophysics, 2002, 81: 887-910. DOI:10.1016/S0074-6142(02)80261-3 Sun Nan, Sun Yaochong, Yu Shan. Numerical Simulation of Seismic Wave Field in medium with limited water[J]. Journal of Seismological Research, 2017, 40(4): 557-564 (in Chinese with English abstract). Wang Baoshan, Ge Hongkui, Wang Bin, Wang Haitao, Zhang Yuansheng, Cai Huiteng, Chen Yong. Practices and advances in exploring the subsurface structure and its temporal evolution with repeatable artificial sources[J]. Earthquake Research in China, 2016, 30(3): 284-297. Wang Hongti, Zhuang Cantao, Xue Bing, Zhao Cuiping, Zhu Zuyang. Precisely and actively seismic monitoring[J]. Chinese Journal of Geophysics, 2009, 52(7): 1808-1815 (in Chinese with English abstract). Yamaoka K., Kunitomo T., Miyakawa K., Kobayashi K., Kumazawa M. A trial for monitoring temporal variation of seismic velocity using an ACROSS system[J]. Island Arc, 2001, 10(3/4): 336-347. Zhang Bo, Bai Chunhua. Research progress on the dynamic characteristics of gaseous detonation[J]. Scientia Sinica Physica, Mechanica & Astronomica, 2014, 44(7): 665-681 (in Chinese with English abstract). Zhang Zhongjie, Deng Yangfan, Teng Jiwen, Wang Chunyong, Gao Rui, Chen Yun, Fan Weiming. An overview of the crustal structure of the Tibetan Plateau after 35 years of deep seismic soundings[J]. Journal of Asian Earth Sciences, 2011, 40(4): 977-989. DOI:10.1016/j.jseaes.2010.03.010