Xin Wng*,Y-n Hu,Sho-ze Luo,Lu-chen Zhng,Bo Wu

aHydraulic Engineering Department,Nanjing Hydraulic Research Institute,Nanjing 210029,China

bState Key Laboratory of Hydrology-Water Resources and Hydraulic Engineering,Nanjing Hydraulic Research Institute,Nanjing 210029,China

cKey Laboratory of Navigation Structure Construction Technology of Ministry of Transport,Nanjing Hydraulic Research Institute,Nanjing 210029,China

Prototype observation and influencing factors of environmental vibration induced by flood discharge

Xin Wanga,b,*,Ya-an Hua,c,Shao-ze Luoa,Lu-chen Zhanga,Bo Wua

aHydraulic Engineering Department,Nanjing Hydraulic Research Institute,Nanjing 210029,China

bState Key Laboratory of Hydrology-Water Resources and Hydraulic Engineering,Nanjing Hydraulic Research Institute,Nanjing 210029,China

cKey Laboratory of Navigation Structure Construction Technology of Ministry of Transport,Nanjing Hydraulic Research Institute,Nanjing 210029,China

Due to a wide range of field vibration problems caused by flood discharge at the Xiangjiaba Hydropower Station,vibration characteristics and in fluencing factors were investigated based on prototype observation.The results indicate that field vibrations caused by flood discharge have distinctive characteristics of constancy,low frequency,small amplitude,and randomness with impact,which signi ficantly differ from the common high-frequency vibration characteristics.Field vibrations have a main frequency of about 0.5-3.0 Hz and the characteristics of long propagation distance and large-scale impact.The vibration of a stilling basin slab runs mainly in the vertical direction.The vibration response of the guide wall perpendicular to the flow is signi ficantly stronger than it is in other directions and decreases linearly downstream along the guide wall.The vibration response of the underground turbine floor is mainly caused by the load of unit operation.Urban environmental vibration has particular distribution characteristics and change patterns,and is greatly affected by discharge,scheduling modes,and geological conditions. Along with the increase of the height of residential buildings,vibration responses show a signi ficant ampli fication effect.The horizontal and vertical vibrations of the 7th floor are,respectively,about 6 times and 1.5 times stronger than the corresponding vibrations of the 1st floor.The vibration of a large-scale chemical plant presents the combined action of flood discharge and working machines.Meanwhile,it is very dif ficult to reduce the low-frequency environmental vibrations.Optimization of the discharge scheduling mode is one of the effective measures of reducing the flow impact loads at present.Choosing reasonable dam sites is crucial.

Flood discharge;Environmental vibration;Vibration characteristics;In fluencing factor;Prototype observation

1.Introduction

Design for flood discharge and energy dissipation in high dams is a crucial part of water conservancy and hydropower engineering construction.The ample energy of high-speed flow needs to be dissipated safely and quickly with a limited amount of energy dissipators.As a result,the flood discharge and energy dissipation design are directly related to the safety of the hydraulic project.In the process of energy dissipation with rolling,shearing,and friction of flow,the energy dissipation structure is constantly under the impact of high-speed flow,and strong vibrations are induced.Many discharge structures have been damaged worldwide.For example,the stilling basin slab of the Sayano-Shushenskaya Hydropower Station in the former Soviet Union was twice destroyed,and the nearly 8 m-thick bottom floor was set off by a huge flow load,shocking the world(Wang and Luo,2012).When the Wuqiangxi Hydropower Station in China was discharging aflood,the large-area stilling basin slab with a 3-m thickness was set off and the scour depth of bedrock was nearly 30 m (Lian,1998).The stilling basin guide walls of the Texarkana Hydropower Station in the USA,the Bapa Zschinsky Hydropower Station in the former Soviet Union,and the Wanan Hydropower Station in China have experienced hydraulic damage(Lian and Ma,2007).In recent years, flow-induced vibration has been a focus in high-speed flow fields,and a great amount of progress has been made in model tests,numerical analysis,and prototype observation after years of continuous research(Wang and Yan,2013;Wang et al.,2009, 2014).A variety of research could lead to the end of severe vibrations of discharge structures.

Most studies on environmental vibrations have concentrated on the field of urban traf fic(Lee and Wang,2012;Xia et al.,2009;Fukada et al.,2012;Prez et al.,2011;Sharp et al.,2014),which is signi ficantly different from flowinduced vibration.Environmental vibrations induced by flood discharge always exist,but not enough attention has been paid to them because hydropower stations are usually far away from urban and rural areas.Therefore,there is a lack of relevant records and cognition about environmental vibrations induced by hydraulic loads of flood discharge.Numerical simulation of water through the radial gates of the Caruchi Dam,in southern Venezuela,and its relation to the vibration of spillways and adjacent control building has been conducted, and the structural vibration source was determined in the case of gate openings of up to 5 m above the normal values (Sanchez and Salazar,2010).Prototype dynamic testing of the left guide wall of the Three Gorges Dam has been carried out under the condition of flood discharge,and the regularity of distribution of the root mean square(RMS)of dynamic responses on the guide wall has been examined(Huang and Li, 2011).The vibration responses of the Qianjiang Tunnel under the impact of the sea bore of the Qiantang River have been monitored and the main frequency of vibration was about 2 Hz (Cai and Huang,2013).The effect of vibrations of the Zhigulevskii hydropower structure on soils in the nearby territories of Tolyatti City(Shumakova and Kotlyakova,2010)and the vibration source and shock absorption scheme of near- field vibrations caused by flood discharge(Yin and Zhang,2014) have been studied.

Recently,a wide range of unexpected environmental vibrations have appeared in Shuifu County near the Xiangjiaba Hydropower Station when flood discharge structures have begun to release floodwater,leading to a certain negative effect on daily production and human livelihoods,because the county is very close to the stilling basin and the shortest direct distance is only 0.5 km.In this study,systematic prototype observation was conducted to investigate the severe environmental vibrations caused by flood discharge.High-accuracy and high-sensitivity vibration sensors(the 941 sensor,made in China)in three directions,specially designed for monitoring of low-frequency vibrations,were employed to measure the vibration displacement,velocity,and acceleration.More than 30 monitoring sites were located within a radius of 2.5 km from the center of the stilling basin.As shown in Fig.1 and Table 1,the monitoring sites were set in the guide wall,stilling basin slab,underground power house,soil surface,residential buildings,and a large-scale chemical plant in the county. Some of these hydraulic structures,such as the guide wall and the underground power house,were built on hard rocks. However,most of the area of the county is artificial backfill and many residential buildings have been built on soft soil foundations.Cases with a series of discharge from 6600 m3/s to 0 m3/s and different discharge scheduling modes were observed.Based on the monitoring data,time domain statistics analysis and frequency domain spectrum analysis procedures such as the fast Fourier transform(FFT)were conducted to describe the vibration characteristics of the vibrations of flood discharge structures,the underground power house,the soil surface,residential buildings,and the large-scale chemical plant in the county caused by flood discharge.Influencing factors and damping measures are proposed.

Fig.1.Location of monitoring sites.

2.Vibration characteristics of hydraulic structures

2.1.Slab of stilling basin

The project's flood energy dissipation works contained two adjacent stilling basins,and a cross-section of the spillway is shown in Fig.2.The stilling basin mainly consisted of a slab,a guide wall,and an end ridge,and there were many galleries inthe bottom plate.Six monitoring sites(I1 through I6 in Fig.1) were set in the grouting gallery,draining gallery,and access gallery of the end ridge to obtain the vibration response of the bottom plate.The vibration process indicated continuous random vibration with impact characteristics.The vertical vibration was stronger than the horizontal vibrations,and vibrationsinthetwohorizontaldirections,paralleland perpendicular to the flow(hereafter simplified as parallel to the flow and perpendicular to the flow)were basically the same.Vibration responses of the grouting gallery and draining gallery near the head of the stilling basin were significantly stronger than vibration responses of the access gallery in the end ridge.Compared with the flow pulsating pressure achieved from model tests,the vibration responses were closely related to pulsating pressures of the hydraulic jump.As the hydraulic jump mainly occurred at the head of the stilling basin where the pulsating pressure was much stronger than it was near the end ridge,the energy of the high-speed flow near the head of the stilling basin was relatively strong and abated quickly downstream.

Table 1 Location and number of monitoring sites.

Fig.2.Cross-section of spillway.

The practically monitored peak values of the vibration displacement,velocity,and acceleration of the grouting gallery and the access gallery in the end ridge when the discharges into the stilling basin are 3360 m3/s and 960 m3/s are listed in Table 2.It was found that vibration responses decreased with the discharge.When one of the two stilling basins was running alone,the vibration of the slab of the running stilling basin was stronger than that of the other stilling basin.The main frequencies of the vibration response of bottom plate were 1.5-2.5 Hz as determined through spectrum analysis of vibration signals.

2.2.Guide wall of stilling basin

The guide wall of the stilling basin is a towering thin-walled structure,and its vibration safety should be given signi ficant attention.Therefore,several monitoring sites(G1 through G4 in Fig.1)were set on the top of the right guide wall to obtain the vibration responses.Under the frequently occurring discharge conditions,the time-history curves of the vibration displacement in three directions are shown in Fig.3.The guide wall presented stable random vibration under the load of dynamic flow.The vibration perpendicular to the flow was significantly stronger than it was in the other two directions. Peak values and RMS values of the vibration displacements, velocities,and accelerations on the top of the guide wall are listed in Table 3.The maximum displacement,velocity,and acceleration of the guide wall were 71.364 μm,1.140 mm/s, and 0.02525 m/s2,respectively.The guide wall presented narrow-band random vibrations and the frequency of the highenergy zone was around 0.5-2.5 Hz as determined through spectrum analysis shown in Fig.4,which was consistent with the energy distribution of the flow pulsating pressure.In order to study the distribution characteristics of vibration along the guide wall,a series of accelerations under similar discharge conditions are shown in Fig.5,and it was found that the vibration decreased linearly along the guide wall.

2.3.Underground power house

In order to evaluate the in fluence of the vibration induced by flood discharge on the turbine units of the underground power house,the vibration responses of the turbine floor were monitored when the dam discharged floodwater.Using adischarge of 2500 m3/s as an example,the vibration responses are listed in Table 4.The turbine floor showed continuous stable periodic vibration with a high main frequency.The acceleration was relatively large and the displacement was relatively small,characteristics that significantly differed from the characteristics of the low-frequency vibration induced by flood discharge.The vertical vibration was slightly stronger than it was in the horizontal direction.The main frequency of vibration(35.7 Hz)was 30 times the rotating frequency of the units(1.19 Hz),whose rotating angular velocity was 71.4 r/min.As a result,the vibration of the underground power house was mainly caused by the unit operation and hardly affected by flood discharge.

Table 2 Peak values of vibration displacement,velocity,and acceleration of grouting gallery and access gallery.

Fig.3.Time-history curves of displacement in different directions on top of guide wall.

Table 3 Vibration responses at site G1 on top of guide wall.

Fig.4.Power spectrum density of displacement in different directions on top of guide wall.

Fig.5.Distribution of guide wall vibration.

Table 4 Vibration responses of turbine floor of underground power house.

3.Environmental vibration characteristics

3.1.Soil surface vibration

The pulsating pressure of the high-speed flow in the stilling basin was the source of environmental vibrations.Downtown was only 0.5 km far from the stilling basin and the vibration phenomenon existed throughout the county.Therefore,more than thirty monitoring sites were set within a radius of 2.5 km from the stilling basin to record the field vibration responses to study the propagation rule of the flow-induced vibration along the foundation.

Using a discharge of 330 m3/s as an example,the contours of the RMS values of displacement and velocity are shown in Fig.6.It was found from vibration distribution that vibration responses did not decrease with the increase of the distance from the source,indicating a special distribution regulation.In the ancient river channel downstream of the right bank,the vibration in the strip area was significantly stronger than it was in other regions.The maximum values of environmental vibration appeared at the site of the ancient river about 1.5 km away from the stilling basin.For example,at a discharge of 6600 m3/s,the peak values of the acceleration perpendicular to the flow,parallel to the flow,and in the vertical direction were, respectively,0.0039 m/s2,0.0038 m/s2,and 0.0025 m/s2;the peak values of velocity in the three directions were,respectively,0.26 mm/s,0.26 mm/s,and 0.12 mm/s;and the peak values of displacement in the three directions were,respectively,15.95 μm,10.59 μm,and 6.30 μm.Vibrations in the two horizontal directions were about the same,and the vibration in the vertical direction was slightly smaller.The main frequency of the vibration response of the soil surface was around 2-3 Hz as determined through spectrum analysis.

Fig.6.Environmental vibration distribution at discharge of 330 m3/s.

3.2.Residential building vibration

In order to study the vibration characteristics of residential buildings and provide data for evaluation of the in fluence of discharge-induced vibration on the daily life of people,many residential buildings were monitored and measuring points were set on the 1st floor,3.5th floor,and37th floor(top)of the buildings.The discharge of 3400 m/s was used as an example.The vibration responses of a residential building with the largest soil surface vibration are listed in Table 5.The time-history curves and power spectrum density of displacements in the three directions of the 7th floor are shown in Fig.7.It was found that the vibration responses increased gradually with the increase of the floor and the horizontal vibrations had a more signi ficant increase compared with the vertical vibration.Vibrations of different floors of all the monitored buildings were counted.The horizontal vibration of the 3.5th floor was about 3 times of that of the 1st floor,and the vertical vibration was only about 1.2 times.The horizontal vibration of the 7th floor was about 6 times of that of the 1st floor,and the vertical vibration was only about 1.5 times.The maximum value of the RMS and peak value of the acceleration at the 7th floor were,respectively,0.0054 m/s2and 0.0198 m/s2;the maximum RMS value and peak value of the velocity at the 7th floor were,respectively,0.38 mm/s and 1.36 mm/s;and the maximum RMS value and peak value of the displacement at the 7th floor were,respectively,27.59 μm and 100.63 μm.The direction of all the extreme vibrationresponses above was perpendicular to flow.Spectrum analysis indicated that the main vibration frequency of the 7th floor was 2-3 Hz,as shown in Fig.8,which was consistent with that of the soil surface.

Table 5 Vibration responses of different floors of one residential building at discharge of 3400 m3/s.

Fig.7.Time-history curves of displacement in different directions on 7th floor.

3.3.Field vibration of large-scale plant

There was a large-scale chemical plant including a lot of very important pipes of chemical devices and precision instruments downtown about 2 km away from the stilling basin. Four monitoring sites(V2,V3,VI3,and VI4 in Fig.1)were set in the plant:three of them in the production area and one in the precision instrument room.The monitoring data showed that the field vibration characteristics in the plant were different from the field vibration characteristics of other soil surfaces.When the production machines were running,the vibration of the plant presented the significant characteristics of a high frequency and small amplitude,mainly caused by many different kinds of machines.There were several highfrequency components of 12.38 Hz,16.63 Hz,and 24.88 Hz. The acceleration was relatively large and the displacement was relatively small.Among the four monitoring sites,vibration response,especially the acceleration in the precision instrument room,was small as it was far away from the production area,but the accelerations at the other three monitoring sites in the production area were large,and the vertical vibration was significantly larger than the horizontal vibrations.It can also be seen from the value,distribution,and main frequency of the displacement that the discharge made some contributions to the field vibration of the plant.When the chemical plant stopped production,the high-frequency loads almost disappeared and vibrations decreased significantly.The acceleration showed the largest decline,especially in the vertical direction,and the velocity took second place.The displacement had a slight drop,and the horizontal displacements almost remained constant,a factor associated with the lowfrequency impact caused by flood discharge.As a result,the large-scale plant vibrations were induced by combined contributions of running machines and flood discharge.

Fig.8.Power spectrum density curves of displacement in different directions on 7th floor.

4.Influencing factors of environmental vibration

4.1.Discharge

Since the dam began to discharge water in October,the discharge has been decreasing gradually.There was a certain relationship between vibration and discharge.Using the monitoring site with the maximum vibration response as an example,the relationships between the RMS values of thevibration acceleration,velocity,and displacement and the discharge,based on a long period of monitoring data,are shown in Fig.9.Meanwhile,fitted relationship formulas can be used to predict the vibration responses at other discharges. In Fig.9,A,V,D,Q,andR2in the formulas are the RMS values of vibration acceleration,velocity,displacement, discharge,and coefficient of determination,respectively.On the whole,vibrations decreased with the discharge,and vibrations decreased by about 90%when the discharge varied from 6600 m3/s to 330 m3/s.In fact,this was because there was a close relationship between vibration and flow pulsating pressure.Usually the flow turbulence energy in the stilling basin decreased as the discharge decreased,i.e.,the pulsating pressure on the stilling basin decreased,which was the reason for the decrease of the environmental vibration response.

4.2.Discharge scheduling mode

Each stilling basin contains five discharge holes which are numbered#1 through#5 from left to right downstream.Under a constant discharge of 960 m3/s,two different scheduling modes were employed:evenly discharging through two holes(#1 and #5)and through four holes(#1,#2,#4,and#5)of the right stilling basin.Using the monitoring site with the maximum vibration response as an example,vibration responses at the site under the two modes are compared in Table 6.The vibration responses were reduced by about 70%when the four-hole mode was used rather than the two-hole mode at the same discharge. Hence,there mustbe a close relationship between environmental vibration and scheduling mode.When the discharge was constant,the vibrations would decrease with an increase of the discharge width ora reduction ofthe discharge perunitwidth.As vibrations were determined by flow pulsation,although the discharge did not vary,the pulsating pressures loaded on the stilling basin caused by discharge under different scheduling modes were significantly different,which caused different environmental vibrations.Hydraulic model tests under the same conditions indicated that the pulsating pressure using the twohole mode was about 2-3 times of that using the four-hole mode.As a result,for discharge of different levels,optimization model tests under different scheduling modes could be conducted to find the best mode that caused the least pulsating pressure to reduce the environmental vibration.Therefore, optimizing the scheduling mode is the most effective vibrationdamping measure for the wide-range environmental vibration induced by flood discharge at present.

4.3.Geological conditions

Overall,environmental vibrations presented obvious regional distribution characteristics,as shown in Fig.6,and vibrations of the ancient river strip area downstream of the right bank were relatively strong.It was clear that there was a close relationship between environmental vibrations and geological structure.Using six monitoring sites within a radius of 1.0 km from the stilling basin crossing the ancient river as examples(IV1 through IV6 in Fig.1),the distributions of RMS velocity in the three directions are listed in Table 7.Vibrations at sites IV5 and IV6 were signi ficantly stronger than those at the other four sites.Because these two sites were located in the ancient river and the thickness of the weak layer was more than 80 m,while the other four sites were on the rock foundation and the thickness of the weak layer was less than 10 m.The vibration on the thick weak layer was about 4-6 times stronger than that on the thin weak layer.

Table 6 Vibration comparison under two different scheduling modes.

Table 7 RMS velocity at different monitoring sites.

5.Conclusions

(1)Field and structural vibrations induced by flood discharge had distinctive characteristics of constancy,low frequency,small amplitude,and randomness with impact.The vibrations presented the main frequencies of 0.5-3.0 Hz,long propagation distance,and large-scale impact,characteristics that were signi ficantly different from those of the usual highfrequency vibration.

(2)The vibration of the stilling basin slab was mainly vertical,and the guide wall vibration response perpendicular to the flow was signi ficantly stronger than it was in other directions and decreased linearly downstream along the guide wall.The vibration response of the underground turbine floor was mainly caused by the load of unit operation,and the impact of flood discharge could be ignored.

(3)The urban environmental vibration had obvious distribution characteristics and change patterns and was greatly affected by discharge,scheduling modes,and geological conditions.Vibration response showed an increasing trend with the increase of discharge.Increasing the discharge width or reducing the discharge per unit width could be very useful to reducing the vibration response.In the ancient river channel downstream of the right bank,the vibration in the strip area was signi ficantly stronger than that in other regions.The strongest vibration was in the ancient river and about 1.5 km away from the stilling basin,where the thickness of the weak layer was more than 80 m,and this vibration was about 4-6 times stronger than that on the thin weak layer at the same distance from the stilling basin.

(4)With the increase of the height of residential buildings, vibration responses had signi ficant ampli fication effects,and there were signi ficant differences between the horizontal and verticaldirections.Horizontaland verticalvibrations ofthe 3.5th floor were,respectively,about 3 times and 1.2 times stronger than the corresponding vibrations of the 1st floor,while those of the 7th floor were,respectively,about 6 times and 1.5 times stronger than the corresponding vibrations of the 1st floor.

(5)The vibration of the large-scale chemical plant presented combined action of flood discharge and working machines.The high-frequency loads of machines made a greater contribution to the acceleration,and the low-frequency loads of flood discharge made a greater contribution to the displacement.

(6)It is very difficult to reduce the wide-range low-frequency vibration,and optimization of the dischargescheduling mode is an effective measure of reducing the impact loads at present.Further research on the effect of the continuous low-frequency environmental vibration on structures and the human body is still necessary.

Cai,Z.X.,Huang,M.S.,2013.Vibration tests on crossing-river tunnel under tidal bore.Chin.J.Geotechnical Eng.35(s2),869-873.http://dx.doi.org/ 1000-4548(2013)S2-0869-05(in Chinese).

Fukada,S.,Usui,K.,Yoshimura,T.,Okada,T.,Hama,H.,Kishi,T.,2012. Effectiveness of dampers in controlling a vibration problem near a highway bridge.J.Civ.Struct.Health Monit.2(2),109-122.http://dx.doi.org/ 10.1007/s13349-012-0022-3.

Huang,J.L.,Li,H.K.,2011.Safety evaluation and feedback analysis of guide wall structure under flood discharge excitation.In:Second International Conference on Mechanic Automation and Control Engineering.IEEE, pp.6064-6067.http://dx.doi.org/10.1109/MACE.2011.5988420.

Lee,C.L.,Wang,Y.P.,2012.Assessment of vibrations induced in factories by automated guided vehicles.Proc.Institution Civ.Eng.Struct.Build. 166(4),182-196.http://dx.doi.org/10.1680/stbu.11.00036.

Lian,J.J.,1998.Study on flow induced vibration of spillway guide wall.J. Hydraulic Eng.29(11),33-37.http://dx.doi.org/10.3321/j.issn:0559-9350.1998.11.007(in Chinese).

Lian,J.J.,Ma,B.,2007.Back analysis algorithm for response of flow-induced vibration in over flow high dam.J.Hydraul.Eng.38(5),575-581.http:// dx.doi.org/10.3321/j.issn:0559-9350.2007.05.010.f(in Chinese).

Sanchez,D.,Salazar,J.E.,2010.On the effects of water discharge through radial gates at the Caruachi Dam.In:ASME 2010 10th Biennial Conference on Engineering Systems Design and Analysis,pp.1-5.http:// dx.doi.org/10.1115/ESDA2010-24080.

Sharp,C.,Woodcock,J.,Sica,G.,Peris,E.,Moorhouse,A.T.,Waddington,D.C., 2014.Exposure-response relationships for annoyance due to freight and passenger railway vibration exposure in residential environments.J.Acoust. Soc.Am.135(1),205-212.http://dx.doi.org/10.1121/1.4836115.

Shumakova,E.M.,Kotlyakova,A.V.,2010.The effect of vibrations in the Zhigulevskii hydropower structure on soils in the nearby territories of Tolyatti City.Water Resour.37(3),306-310.http://dx.doi.org/10.1134/ S009780781003005X.

Wang,X.,Li,T.C.,Zhao,L.H.,2009.Vibration analysis of large bulb tubular pump house under pressure pulsations.Water Sci.Eng.2(1),86-94.http:// dx.doi.org/10.3882/j.issn.1674-2370.2009.01.008.

Wang,X.,Luo,S.Z.,2012.Flow-induced vibration study of tunnel spillway working gate on one reservoir.Appl.Mech.Mater.16(13),226-228. http://dx.doi.org/10.4028/www.scienti fic.net/AMM.226-228.13.

Wang,X.,Yan,X.J.,2013.Dynamic optimization and flow-induced vibration study on plate valve of ship lock.Port Waterw.Eng.12,151-154.http:// dx.doi.org/10.3969/j.issn.1002-4972.2013.12.029(in Chinese).

Wang,X.,Luo,S.Z.,Liu,G.S.,Zhang,L.C.,Wang,Y.,2014.Abrasion test of flexible protective materials on hydraulic structures.Water Sci.Eng.7(1), 106-116.http://dx.doi.org/10.3882/j.issn.1674-2370.2014.01.011.

Xia,H.,Chen,J.G.,Wei,P.B.,Xia,C.Y.,2009.Experimental investigation of railway train-induced vibrations of surrounding ground and a nearby multistory building.Earthq.Eng.Eng.Vib.8(1),137-148.http://dx.doi.org/ 10.1007/s11803-009-8101-0.

Yin,R.G.,Zhang,J.H.,2014.Vibration source and shock absorption scheme research of near- field vibration caused by flood discharge and energy dissipation of a stilling pool.Rock Mech.Appl.Civ.Min.Petroleum Eng. 107-116.http://dx.doi.org/10.1061/9780784413395.013.

Received 28 January 2016;accepted 11 November 2016

Available online 9 March 2017

This work was supported by the National Natural Science Foundation of China(Grants No.51479124 and 51109143),the Open Cooperation Fund of State Key Laboratory of Hydraulics and Mountain River Engineering(Grant No.SKHL1422),and the Nanjing Hydraulic Research Institute Foundation (Grant No.Y115006).

*Corresponding author.

E-mail address:xwang@nhri.cn(Xin Wang).

Peer review under responsibility of Hohai University.

http://dx.doi.org/10.1016/j.wse.2017.03.001

1674-2370/?2017 Hohai University.Production and hosting by Elsevier B.V.This is an open access article under the CC BY-NC-ND license(http:// creativecommons.org/licenses/by-nc-nd/4.0/).

?2017 Hohai University.Production and hosting by Elsevier B.V.This is an open access article under the CC BY-NC-ND license(http:// creativecommons.org/licenses/by-nc-nd/4.0/).