NON-LINEAR RESPONSE OF A CONCRETE GRAVITY DAM SUBJECTED TO NEAR-FIELD GROUND MOTIONS

Transcription

1 NON-LINEAR RESPONSE OF A CONCRETE GRAVITY DAM SUBJECTED TO NEAR-FIELD GROUND MOTIONS by K. Jagan Mohan, Pradeep Kumar Ramancharla in 15th Symposium on Earthquake Engineering (15SEE) Report No: IIIT/TR/2014/-1 Centre for Earthquake Engineering International Institute of Information Technology Hyderabad , INDIA December 2014

2 974 15th SEE-2014 NON-LINEAR RESPONSE OF A CONCRETE GRAVITY DAM SUBJECTED TO NEAR-FIELD GROUND MOTIONS K. Jagan Mohan 1 and Ramancharla Pradeep Kumar 2 ABSTRACT Lifeline structures like dams should always be designed for highest safety, resisting worst forces of nature. However, events that took place across the world proved how devastating an earthquake could be, particularly in the near-field areas. Landers earthquake (1992 Mw 7.3), Northridge earthquake (1994 Mw 6.7), Hyogoken-Nanbu earthquake (1994 M w 6.8) and few more earthquake events caused damage due to near-field effects. Notable damage to Koyna dam due to Koyna earthquake (1967 M w 6.5), India and Shih-Kang DamdueChi-Chi earthquake (1999 M W 7.6), Taiwan proved earthquakes and near-field effects can cause damage even to dams. In India there are several dams constructed near the active faults, which might suffer moderate to severe damage in an event of earthquake in near-field areas. In this concern, a proper study on behaviour of concrete gravity dam subjected to near-field ground motions should be performed. In the proposed study, a concrete gravity dam is selected from the National Importance Dams of India (NRLD-2009) and numerically modeled using Applied Element Method (AEM). Fault normal and fault parallel components of 5 near-field ground motion are considered as input. The behaviour of dam subjected to near-field ground motions is then studied by understanding the displacement response, and stress distribution on upstream and downstream side of the dam body. INTRODUCTION Numerous failures to civil engineering structures were observed in the near-field of the 1994 Northridge, California earthquake and the 1995 Hyogoken-Nanbu, Japan earthquake. Among different structures, the consequences of a large dam failing can be disastrous. Although concrete dams have performed extremely well in the past, few lessons were learned on or after summarizing the performance of 19 concrete damsthat have been shaken by Peak Horizontal Ground Accelerations (PHGA) exceeding 0.3g during earthquakes worldwide in past 100 years [Larry, 2012]. However due to 1999 M7.6 Chi-Chi earthquake, Shih Kang dam in Taiwan reported to have failed, where the 1 Asst. Professor, Mahatma Gandhi Institute of Technology, Hyderabad, India jaganmohan@mgit.ac.in 2 Professor, Earthquake Engineering Research Centre, IIIT-H, Hyderabad, India ramancharla@iiit.ac.in

3 K. Jagan Mohan and Ramancharla Pradeep Kumar 975 dam located directly over a fault had ruptured and caused a vertical differential movement of about 9m and horizontally by 2m [Larry, 2012]. Regardless of what has happened, the possibility of a large differential movement occurring in a fault traversing the dam foundation shall not be considered as it is a rare case and should be avoided by a suitable geological study. Other than this one example, in the epicentral area of an earthquake a number of concrete gravity dams have experienced ground shaking and suffered minor damages. Pacoima dam located in Southern California, USA has been shaken by two major earthquakes, the arch dam survived well with the main damage being 2 inches opening of the contraction joint between the arch dam and the thrust block at the left abutment [Larry, 2012]. In this study, the effects of ground motions recorded in the near-field on the seismic performance of concrete gravity dam are investigated by using Applied Element Method [Hatem, 1998]. The nonlinear analysis of the dam is performed by using tension and compression concrete Maekawa Model [Hatem, 1998]. In all analyses, displacements over the height of the dam, tensile stresses on upstream and downstream side of the dam are studied for all near-field ground motions. NUMERICAL METHOD Applied Element Method (AEM) [Meguro et al., 1997 &2000 and Hatem, 1998], is a discrete element method for structural analysis and is used for this study. In AEM, the structure is divided virtually and modeled as an assemblage of relatively small elements. A set of normal and shear springs (Fig. 1a) are distributed along the element faces to connect the contact between elements. The governing dynamic equation for a structure used in AEM is Mu + Cu + Ku = f ( t) Mu (1) g Where, [M] is mass matrix; [C] is damping matrix; [K] is nonlinear stiffness matrix; f(t) is incremental applied load vector u and is derivatives are the incremental displacement, velocity and acceleration vectors respectively. The above equation is used in AEM and is solved numerically using Newmark s beta method. Each element is having three degrees of freedomin 2 dimensional models (Fig. 1b).The general stiffness matrix components corresponding to each degree of freedom are determined by elemental mass and mass moment of inertia are assumed lumped at the element centroid so that it will act as continuous system, and damping matrix is calculated from first mode. The material used in this analysis is Maekawa compression model [Hatem, 1998]. Finally, stresses and strains are defined based on the displacements of the spring end points. NUMERICAL APPLICATION Koyna Dam The Koyna Hydroelectric Project is the largest completed hydroelectric power plant in the state of Maharashtra, India. It is a complex project consisting of total four dams. The tallest monolith of the concrete gravity section of the Koyna dam located on the Koyna River is selected for numerical application. It is also one of the National Importance Dams among 5,100 large dams in India [National Register of Large Dams 2009]. This particular dam previously has experienced one M w 6.5 earthquake in It has also experienced 17 earthquakes of M 5.0 and over 150 earthquakes of M 4.0 [Gupta et.al. 2002]. The seismic safety of 103m high concrete gravity Koyna dam located

4 976 15th SEE-2014 in highly seismic zone had been widely debated and because of that probably this dam has been studied very extensively and so considered for the study. Fig. 2(a) describes the seismic zonation map of India mapped with 1040 active faults and lineaments and 69 National Importance Dams [NRLD-2009]. Numerical Modeling of non-overflow section of Koyna dam The structure is 103m high, 70m and 15m wide at the base and crest respectively. The dam on the downstream is planned and designed out of conventional style by providing large neck portion in an inclined position. The change of cross section of dam at the neck happens at 67m height from the base (Fig. 2(b)). Information regarding the characteristics of the dam was considered from past research [Rajibet. al., 2007] and presented in table 1. Material a) Element formulation in AEM b) Spring connectivity Modulus of elasticity (E c ) (kn/m 2 ) Table 1: Material properties of Koyna dam Poisson ratio (υ) Mass (ρ) (t/m 3 ) Tension Resistance (σ t ) (kn/m 2 ) Compression Resistance (σ c ) (kn/m 2 ) Dam E E E04 Earthquake ground motions Fig. 1: Element modeling in AEM The effects of near-field ground motion on the seismic performance of Koyna dam should be investigated because of three significant faults (Chiplun, West Coast, and Warna fault) close to dam site. Fault normal and fault parallel ground motion components of 1978 Tabas, 1989 Loma Prieta, 1992 Landers, 1994 Northridge, 1995 Kobe earthquakes are used in the analyses. These suites of records were used in system analyses for the SAC Steel Project funded by Federal Emergency Management Agency [NISEE, 2000]. Table 2 provides detailed information on the near-field records used. The time histories for acceleration and velocity of records are presented in Fig. 3. The response spectra of these ground motions for 5% damping ratio with IS-1893 (Part 1): 2002 design spectra for rock site are presented in Fig. 4.

5 K. Jagan Mohan and Ramancharla Pradeep Kumar 977 (a) Fig. 2: (a) seismic Zonation, Faults and National Importance Dams of India [Mohan, 2013], (b) Cross section of Koyna gravity dam (b) Fig. 3: Time histories of acceleration and corresponding velocity of both FN and FP

6 978 15th SEE-2014 Table 2: Near-field ground motion records selected for analyses [NISEE, 2000] Earthquake record Type PGA Magnitude Epicentral distance (km) 1978, Tabas (NF01) FN , Tabas (NF02) FP , Loma Prieta (NF03) FN , Loma Prieta (NF04) FP , Landers (NF11) FN , Landers (NF12) FP , Northridge (NF13) FN , Northridge (NF14) FP , Kobe (NF17) FN , Kobe (NF17) FN *FN = Fault Normal; FP = Fault Parallel NUMERICAL RESULTS AND DISCUSSIONS Dynamic characteristics of dam body Modal analysis has been performed to obtain the dynamic characteristics of dam body. The first nine mode shapes and corresponding natural frequencies are shown in Fig.5. The fundamental periods of the dam for nine lowest modes vary from to s, and are falling in ascending region of the response spectra of few components of earthquakes as shown in Fig. 4. Nonlinear time-history analyses The nonlinear time-history analyses of Koyna dam are conducted by using AEM [Hatem, 1998] for entire duration of the records. However, Trifunac and Bradyduration [Kramer, 2008] of all ground motions is considered to understand the effect of duration on behavior of dam. Fig. 4: Response Spectra of (a) fault normal and (b) fault parallel acceleration components for 5% damping of 5 near-field ground motions and IS 1893 (Part1): 2002 design spectrum

7 K. Jagan Mohan and Ramancharla Pradeep Kumar 979 Fig. 5: First nine modes and their natural frequencies Displacements The absolute maximum horizontal displacements at crest of the dam subjected the components of near-field ground motions are compared in Table 3. The presence of directivity effects with peak amplitudes occurring at the very beginning of the record increases the displacement at the crest of dam. This is observed for FN and FP components of Northridge earthquake. Displacement response at the crest of dam for entire ground motion is given in Fig.6. The absolute maximum horizontal displacements over the height of the dam are given in Fig. 7. It is seen from Table 3 and Fig. 6 that the horizontal displacements vary depending upon the characteristics of ground motions. Principal stresses The maximum principal compressive stresses obtained from the nonlinear time-history analyses for both components of near-field ground motions are lower than the compressive strength of concrete used in dam body. The maximum principal tensile stresses on upstream and downstream face of dam are given in Fig. 8. On the upstream at the heel,moderate tensile stresses are observed for both FN and FP components of all ground motions and maximum stresses are found to be at neck portion (67m), where there is change in cross section of the dam. This has been observed for FN and FP components of Northridge and only for FP component of Kobe acceleration records (Fig. 8a, 8c). On the downstream side at the toe, zero tensile stresses are observed and has gradually increased up to 67 m and then reduced to minimum at the crest for all ground motions and for both FN and FP components. However, maximum tensile stresses are witnessed for Northridge and Kobe ground

8 980 15th SEE-2014 Table 3: The absolute maximum horizontal displacements at crest of the dam Displacement (mm) Fault Normal Fault Parallel Tabas Loma Prieta Landers Northridge Kobe Fig. 6: Displacement response at the crest of the dam for FN and FP components of 5 near-field ground motions Fig. 7: Displacement response over the height of the dam

9 K. Jagan Mohan and Ramancharla Pradeep Kumar 981 motion records (Fig. 8b, 8d) at change in cross section of the dam. The contours of Koyna dam subjected to FN and FP components of Northridge and Kobe ground motions are given in Fig. 9. Fig. 8: Maximum principal tensile stress on upstream and downstream sides of the Koyna dam, analyzed for FN and FP components of 5 near-field ground motions (a) Northridge FN (b) Northridge FP (c) Kobe FN (d) Kobe FP Fig. 9: Maximum tensile stress [MPa] contours of Koyna dam for FN and FP components of Northridge and Kobe ground motions

10 982 15th SEE-2014 It is seen from the tensile stress contours that the stress concentration occurs at the center portion of the neck, where the cross-section of dam changes. Moderate stresses are observed on the downstream side of the dam. CONCLUSIONS The investigations of the seismic performance of concrete gravity dams to near-field ground motions taking into account its components has led to the following conclusions: Displacements over the height of the structure are observed to change at 67m from base of dam for Northridge and Kobe ground motions and maximum observed displacements at the crest are for Northridge ground motion.this is because of high amplitude long period velocity pulse (1.5 s) of motion occurring at the beginning of the record (between s). Such a long period pulse of motion with high amplitude has not been seen in any other ground motion. The nonlinear time-history analyses introduce that the stress concentration occurs at the centre portion at the change in cross-section of dam for Northridge and Kobe ground motions, forcing the dam to fail in tension. This study confirms the importance of the ground motion selection for appropriate evaluation of the performance of concrete gravity dams. REFERENCES 1. Alemdar Bayraktar, Temel Türker, Mehmet Akköse, Şevket Ateş, The effect of reservoir length on seismic performance of gravity dams to near- and far-fault ground motions, Natural Hazards, Volume 52, Issue 2, pp , February Fabio Mazza, Alfonso Vulcano, Nonlinear dynamic response of R.C. framed structures subjected to near-fault ground motions, Bulletin of Earthquake Engineering, Volume 8, Issue 6, December Haibo Wang, MinghuiFeng, Huichen Yang, Seismic nonlinear analyses of a concrete gravity dam with 3D full dam model, Bulletin of Earthquake Engineering, Volume 10, Issue 6, pp , December Rajib Sarkar, D K Paul, L Stempniewski, Influence of Reservoir and Foundation on the Non-linear Dynamic Response of Concrete Gravity Dam, ISET Journal of Earthquake technology, Volume 44, Number 2, pp , June SüleymanAdanur, Ahmet Can Altunişik, AlemdarBayraktar, Mehmet Akköse, Comparison of near-fault and far-fault ground motion effects on geometrically nonlinear earthquake behavior of suspension bridges, Natural Hazards, Volume 64, Issue 1, pp , October Tatsuo Ohmachi, AbdolrahimJalali., Fundamental Study on Near-Field Effects on Earthquake Response of Arch Dams, Earthquake Engineering and Engineering Seismology, Volume 1, Number 1, pp. 1 11, September Tatsuo Ohmachi, Naoyuki Kojima, Atsushi Murakami, Nobuhiko Komaba, Near-Field Effects of Hidden Seismic Faulting on a Concrete Dam, Journal of Natural Disaster Science, Volume 25, Number 1, pp. 7-15, Books 8. amalesh Kumar, a text book on Basic Geo-technical Earthquake Engineering, Tagel-Din-Hatem, A New Efficient Method for Nonlinear, Large Deformation and Collapse analysis of Structures, PhD Thesis (1998), The University of Tokyo, Japan.

11 K. Jagan Mohan and Ramancharla Pradeep Kumar 983 Conference Proceedings 10. Abdolrahim Jalali, Tatsuo Ohmachi, Aspects of Concrete Dams Response to Near-Field Ground Motions, The 12 th World Conference on Earthquake Engineering, Larry K Nuss, Norihisa Matsumoto, Kenneth D Hansen, Shaken but not Stirred Earthquake Performance of Concrete Dams, USSD Proceedings, Paul G Somerville, Engineering Characterization of Near-Fault Ground Motions, NZSEE conference, R Pradeep Kumar, K Meguro, Applied Element Simulation of non-linear Behaviour of Dip-Slip Faults for Studying Ground Surface Deformations, Seismic Fault-induced Failures, pp , January 2001.