Earthquake Reaearch in China  2018, Vol. 32 Issue (4): 584-601
Seismic Damage to Owner-Built RC Frames in Charikot during the 2015 Nepal Earthquake Sequence
Qu Zhe, Wang Tao, Lin Xuchuan, Zhang Haoyu, Yang Yongqiang
Key Laboratory of Earthquake Engineering and Engineering Vibration, Institute of Engineering Mechanics, CEA, Sanhe 065201, Hebei, China
Abstract: The damage to the masonry-infilled reinforced concrete (RC) frame buildings in Charikot, the capital city of Dolakha district in Nepal, during the 2015 April-to-May Nepal earthquake sequence is reported. Most of these buildings were built by the owners with little governmental inspections regarding their structural design or constructional quality. Although they generally performed better than other structural systems such as stone-masonry houses, the RC frames sustained extensive damage ranging from cracking of infill to complete collapse. In particular, eight of the 72 inspected RC frames alongside an uphill street collapsed in different ways. In addition to the un-engineered nature of these RC frames, their collapse could also be attributed to multiple technical reasons including the effect of terrain, the pounding between adjacent buildings and the accumulative damage in the earthquake sequence.
Key words: Nepal earthquake     Reinforced concrete frame     Masonry infill     Slope     Pounding     Accumulated damage

INTRODUCTION

Reinforced concrete (RC) moment resisting frames are widely used as a ductile lateral system for modern multi-story and high-rise buildings. In addition to the open space they provide for architectural flexibility, they are deemed to be an efficient structural system for their high ductility under earthquake excitations. However, the extensive damage to RC frames in historic major earthquakes, such as the 1985 Mexico City earthquake (Villaverde, 1991), the 1995 Kobe earthquake (AIJ, 1997), the 2008 MS8.0 Wenchuan earthquake (Ye et al., 2008), and the MS9.0 Tohoku earthquake (AIJ, 2012) to name only a few, revealed inherent weaknesses, among which are the weak story collapse (also known as pancake collapse) and other typical damage including brittle shear failure of columns and premature failure of beam-to-column joints.

Before these problems were properly addressed, RC frames have been introduced to Nepal as an earthquake-resistant system to replace vernacular construction, because of the relatively low cost of RC frames as compared to other modern systems such as RC wall structures and steel construction. Take the Kathmandu Valley where the national capital is located as an example. As per the census data, the population in the Kathmandu Valley, including the Kathmandu, Lalitpur and Bhaktapur districts, increased by 77% from 1.39 to 2.46 million in the two decades between 1991 and 2001. During the same period of time, the number of households increased from 256, 203 to 613, 606, a 140% increase. Along with the rapid growth of the population, the proportion of RC frames in the total building inventory in the valley increased from 23% in 1991 to 37% in 2011 (JICA, 2002a, HDX, 2015). Today, RC frames can be seen all over Nepal even in remote mountainous villages where the transportation of the construction material is extremely difficult.

While vernacular constructions such as stone masonry and timber structures are receiving less attention, RC frames in Nepal are becoming a modern vernacular construction (Hicyilmaz, 2015), which are usually built by the owners and local craftsmen following some rules of thumb which are insufficient to ensure the seismic performance of buildings, and to make it worse, construction quality is usually poor (Dixit, 2004). By numerical analyses of the seismic performance of a typical RC frame with four different designs, Chaulagain et al. (2013) show that the current construction practice (CCP) in Nepal provides much lower earthquake-resistant capacity for RC frames than those guaranteed by the Nepalese or Indian building codes. The example building used for the study is a 2 by 3 span 3-story frame without masonry infills. In the CCP case, all columns have 230 by 230mm cross sections reinforced by 4 or 6 longitudinal bars 10mm in diameter; all beams have 325 by 230mm cross sections with 2 or 3 longitudinal bars 12 mm in diameter on both the top and bottom of the section. As will be shown later, this practice may have evolved in the past few years and may differ among areas. In addition to the experience-based proportioning of the structural members, field investigations show that the construction process of the owner-built RC frames is also quite arbitrary. In some cases, the owner-built frames were constructed in the same way as well-engineered frames, that is, the RC columns and beams were cast before masonry walls were built. In some other cases, however, the masonry infills were built before casting the beams so that the top edge of the masonry can be used as the formwork of the beam. This makes the system closer to a confined masonry structure, but the masonry edges are not staggered for a better bond with the concrete. There were also cases in which the masonry walls in the first story were built before the beam, whereas those in the second story were built after the beams and slabs were cast. As a result, it is difficult to tell how much gravity load is carried by the masonry walls in an owner-built RC frame.

The deadly Nepal earthquake sequence in April and May, 2015, provided a rare chance to examine the seismic performance of the owner-built RC frames in Nepal. As part of a wider-scope reconnaissance project (Sun and Yan, 2015), a detailed inspection for the owner-built RC frames was conducted in Charikot, the headquarters of Dolakha district. Reconnaissance efforts of various scales would provide invaluable lessons that can help improve the seismic performance of buildings. Detailed and comprehensive investigations on a specific building, which usually include both field inspections and numerical simulations, are useful to identify the reasons why or why it wasn't damaged. The findings on the specific building can then be generalized to propose improvements for structures of the same type (Mahin and Bertero, 1975, Tuna and Wallace, 2015). However, such a generalization may sometimes lead to erroneous conclusions if the limitations of the findings on a specific building at a specific site are not fully recognized. In some cases, the studied 'typical' building is not really typical from a statistical point of view, partly because people usually focus more on damaged buildings rather than intact ones during field inspections. To avoid such a bias, damage assessments were conducted on every single RC frame along a selected street in Charikot, which makes a statistical interpretation of the extent of the damage possible.

1 THE NEPAL EARTHQUAKE SEQUENCE IN APRIL-MAY, 2015

In April and May, 2015, a series of major earthquakes, including the MS8.0 main shock on April 25 and three aftershocks with a surface wave magnitude of no less than 7.0, hit western and central Nepal (Fig. 1). They caused tremendous loss of lives and extensive damage to buildings. According to the Ministry of Home Affairs, Nepal, the earthquake sequence killed 8, 898 people, destroyed 605, 257 houses and partially damaged 285, 099 houses as of July 29, 2015 (UNOCHA, 2015). Nepal has long been identified as one of the regions with the highest seismic risk in the world. In a report by JICA (2002a), an earthquake scenario named the Mid Nepal Earthquake, which is very similar to the April 25 main shock in both magnitude and location, was anticipated. The report describes a huge earthquake of M8.0 on an earthquake fault in the middle of Nepal that may be followed by M7.0 aftershocks. The report also predicts that about 18, 000 people will be killed in the Kathmandu Valley alone by such an earthquake. Fortunately, the death toll in the Kathmandu Valley during the 2015 earthquake sequence was approximately 1800 (UNHCR, 2015), much less than the JICA prediction, and the function of Kathmandu as the capital city of Nepal recovered more rapidly than expected.

 Fig. 1 Locations of epicenters of 2015 earthquake sequence

As part of the aid project for the Nepal earthquake launched by the Chinese government, a multidisciplinary reconnaissance team of 22 experts was dispatched by the China Earthquake Administration (CEA) to Nepal about forty days after the main shock to evaluate the seismic damage and losses. During their two week visit, the six districts in the central northern Nepal as underlined in Fig. 2 were visited for building damage inspections. The death toll for each district is also provided under the name of the district in Fig. 2 (UNHCR, 2015). A general review of the on-site building damage observed by the CEA team can be found in Sun and Yan (2015), in which the commonly-used structural systems for buildings in Nepal are classified into five types, namely reinforced concrete (RC) frame structures, stone masonry structures, brick-wood structures, adobe structures, and historical structures.

 Fig. 2 Districts investigated by CEA team after the earthquakes (Figures in brackets is death toll by district as of June 3, 2015)
2 SEISMIC DAMAGE TO RC FRAMES IN CHARIKOT, DOLAKHA 2.1 Inspected Region

The inspected region concerned in this paper is located in Charikot, the headquarters of Dolakha district. Also known as Bhimeshwar, the city is located at an altitude of 1, 554m among the mountains. As shown in Fig. 2, it is approximately 150km southeast to the epicenter of the MS8.0 main shock, but is within 30km south of the epicenter of the MS7.5 aftershock on May 12.

According to the 2011 Nepal census data, the municipality has a population of 22, 537 and 6, 076 households. 775 households (12.7%) lived in RC buildings whereas the majority of 4, 462 households (73.4%) lived in masonry structures with mud mortar (HDX, 2015). Although constituting only a small fraction of the municipal building inventory, RC buildings predominate in the downtown area, where six streets radiate from a small square (Fig. 3). From the square, some of the streets go down to the valley, while two of them go up to the mountain ridges.

 Fig. 3 Downtown of Charikot municipality: inspected and collapsed RC framed buildings

Eleven RC framed buildings collapsed in the earthquake sequence within the area shown in Fig. 3. Their locations and photos are also included in Fig. 3. Some of these buildings exhibited obvious weak story mechanisms at the lower stories (e.g., Building A, B, C, G and K), whereas some others collapsed at the top stories (e.g., Building H and I). Some completely fell on the ground (e.g., Building D, F and J) while some others sustained partial collapse due to the impact from neighboring buildings that collapsed (e.g., Building E knocked by Buildings D and F).

2.2 Inspected RC Buildings

A detailed investigation was conducted in the area along one of the uphill streets shown in Fig. 3 and Fig. 4. A total of 72 RC framed buildings were inspected for seismic damage. Most of these buildings were occupied as residence or hotels before the earthquakes. As shown in Fig. 5, three- or four-story buildings with two or three bays in both directions were most commonly seen.

 Fig. 4 East view of inspected uphill streets with several collapsed buildings identified

 Fig. 5 Statistics of (a) number of stories and (b) number of spans of RC framed buildings in the inspected region

Typical structural layouts of the RC frames in the inspected area are depicted in Fig. 6. The RC frames were usually infilled with brick masonry walls. Most buildings had cast-in-situ concrete slabs of 100mm thick. The floor slabs and the supporting RC beams usually extended outside the columns to form balconies over the street. The story height, H, is usually uniform among stories and is 2.8m in most cases. The mid-span is usually used as corridors and is much shorter than the side spans. The statistics of eight carefully inspected buildings shows an average span width of 1.7m for the mid spans and 3.2m for the side spans along the street (i.e., x-direction in Fig. 6(a)) (Fig. 7(a)). In some cases, the RC columns along one side of the corridor were eliminated probably to reduce the cost (Fig. 6(b)). This resulted in two-span frames of uneven span widths, although they appeared to be three-span buildings. In the direction transverse to the street, the average span width is even larger (3.9m). The average column depth and width are both approximately 280mm and the average cross sectional reinforcement ratio is approximately 1.3% (See Fig. 8(a) for reference). The average beam depths and width are approximately 350mm and 260mm, respectively.

 Fig. 6 Typical configurations of RC framed buildings in Charikot

 Fig. 7 Statistics of (a) span lengths, L, and (b) cubic concrete compressive strength, fcu, of selected RC framed buildings in the inspected region

 Fig. 8 Cross sections of (a) columns and (b) beams in the example RC frame

Schmidt hammer tests were performed for the concrete compressive strength of five of the inspected buildings. The estimated 150-mm cubic strengths, fcu, ranged from 17.6 to 29.5MPa with an average of 25.9MPa. It is higher than the specified minimum concrete cubic strength of 20MPa in the Nepalese guideline for owner-built low-rise RC buildings (NBC 205, 2012). The fitted normal distribution of the concrete strength is given in Fig. 7(b).

To give a general idea of the earthquake resistance of the RC frames inspected in Charikot, the three-span four-story frame in Fig. 6(a) is taken as an example. The aforementioned average values are assumed for the dimensions of the structure, that is, story height, H=2.8m; side- and mid-span lengths along street, Lx1=3.2m, Lx2=1.7m, span length transverse to the street, Ly1=Ly2=3.9m. The average cross sections as shown in Fig. 8 are assumed for all columns and beams. The concrete cylinder compressive strength, fc', is taken as 0.8fcu=20.5MPa according to the field test. The yield strength of steel rebar is assumed to be 415MPa per the Nepalese code for RC construction (NBC 205, 2012). The averaged beam and column sections in Fig. 8 are very close to those stipulated NBC 205 (2012), that is, 355mm deep and 230mm wide beams with 125mm thick slabs, and 300 by 300mm columns with four 16mm diameter and four 12mm diameter longitudinal bars. However, it should be noted that these are for low-rise buildings of a maximum height of 11m or three stories in NBC 205 (2012).

In estimating the weight of the structure, the three exterior bays of the frame that do not face to the street and half of the interior bays are assumed to be infilled with 150mm thick brick masonry walls. As a result, the dead load including the columns and beams, the slab and the masonry infills, is approximately D=651kN on each story assuming the mass densities of reinforced concrete and brick masonry are 25kN/m3 and 18kN/m3, respectively. Further considering a live load of 2kN/m2 on each floor excluding the balcony, which is a commonly used value for residential buildings in China, the combined weight of a story, D + 0.5L=746kN. This gives the total seismic weight of the four-story building, W=3, 506kN.

By assuming a column hinge mechanism at the bottom story, the base shear capacity, Vp, of the frame can be approximately estimated through plastic analysis, that is, Vp=2nMpc/H, where n=16 is the number of columns; Mpc is the column plastic moment capacity, which is a function of the axial force in the column. The axial force on each column is estimated by the combined dead and live load without considering the additional axial force imposed by overturning moment. According to the sectional analysis in Response2000, the column flexural capacity, Mpc, at the bottom story varies from 54.0 to 59.7kNm depending on their axial force, which ranges from 180 to 245kN (0.12 to 0.16 axial force ratio). The corresponding base shear capacity, Vp, under a column hinge mechanism is 637kN. The resultant base shear coefficient for the example building is therefore Vp/W=0.18. It should be noted that this coefficient corresponds to the formation of a full plastic mechanism. It can be directly compared to the required base shear coefficient in the Level Ⅱ seismic design in the Japanese practice (BSL, 2000) (Table 1). In many countries including Nepal, however, the strength demand is calculated at the elastic limit state. If taking into account an over-strength factor of 2.0 to represent system redundancy, the base shear coefficient corresponding to the elastic limit state, Ve/W, becomes 0.09. It still satisfies the minimum requirement of 0.09 by NBC 205 (2012), yet falls below the code requirements for RC moment-resisting frames in some other earthquake-prone areas as listed in Table 1.

Table 1 Base shear coefficients for short-period RC frame buildings

The above comparison is only to provide a general idea of the seismic resistance of the inspected RC buildings in Charikot. The strength estimated by the plastic analysis represents an upper bound solution for the real strength. The assumed over-strength factor that reduces the plastic strength to the elastic limit is arbitrary. In addition, the actual strengths of the inspected buildings are subjected to large uncertainties from multiple sources includes, but is not limited to the various structural configurations, the quality of the construction materials and the existence of masonry infills.

2.3 Damage Statistics

Although the inspected RC frames don't seem as fragile as expected on an average account, extensive damage including complete collapse was observed (Fig. 9). The damage of all the 72 RC frame buildings along the inspected uphill street is classified into five levels, namely slight damage, minor damage, moderate damage, severe damage, and collapse. The resultant damage levels are depicted in Fig. 10. The damage level classification conforms to the Chinese national standard ─ Classification of Earthquake Damage to Buildings and Special Structures (GB/T 24335-2009). The underlying concept of the classification is to correlate the damage levels to the functionality of the damaged buildings. Visual damage to both structural and non-structural components (i.e., the masonry infills) is taken into account in the determination of the damage level. Visual damage to RC beams and columns includes cracks, concrete spalling, rebar exposure and rebar buckling. For masonry infills, this includes cracks and out-of-plane deformation (see Tables 2 and 3 for more details).

 Fig. 9 South view along the inspected uphill street showing collapsed buildings: Building A on left and K on right

 Fig. 10 Damage levels of inspected RC framed buildings in Charikot

Table 2 Classification of damage levels of reinforced concrete framed buildings in GB/T 24335-2009

Table 3 Classification of cracks based on crack widths

Although efforts were made to keep the classification standard as quantitative and objective as possible, professional experience and engineering judgement play a critical role in the on-site practice. Besides, the standard does not include foundation damage, which may impose a critical concern for the reparability of a building. In classifying the damage levels of the inspected buildings in Charikot, the foundation damage, which was observed in seven of the buildings, was also taken into account by engineering judgement.

As shown in Fig. 11, 60% of inspected buildings remained intact or sustained only slight to minor damage, whereas 11% collapsed (eight buildings), 12% were severely damaged, and 17% moderately damaged. A damage index, d, which takes 0 for slight damage and 1 for collapse, is assigned to each damage level in Table 1. The weighted average of the damage indices of multiple buildings, referred to as Da, can be used to represent the overall extent of their damage (equation 1).

 Fig. 11 Damage level statistics of inspected RC frame buildings in Charikot
 ${D_{\rm{a}}} = \mathop \sum \limits_i {d_{\rm{i}}}{\lambda _{\rm{i}}},$ (1)

where di is the damage index associated with the ith damage level, 0 for no damage and 1 for collapse; li is the percentage of buildings whose damage is classified into the ith damage level.

With the assumed di values in Table 1 and the li data for each damage level as shown in Fig. 11, the average damage index, Da, of the 72 inspected buildings is 0.35.

Among other structural properties, the base shear capacity of an RC framed building is most relevant to its damage. For the owner-built buildings that lack a seismic design procedure to keep an explicit minimum standard, the base shear capacity is directly related to the number of stories of the building if the plan layouts and components properties are similar. In Fig. 12(a), the significant difference in the damage levels between low-rise (3 stories or less) and medium-rise (more than 3 stories) buildings is shown. In particular, 35% medium-rise buildings collapsed or severely damaged, whereas this ratio for low-rise buildings is less than 7%. The average damage index, Da=0.46, of the medium-rise buildings is also much larger than that of the low-rise buildings.

 Fig. 12 Relationship between damage level of the inspected buildings and (a) number of stories and (b) number of spans

The number of spans may also have a potential influence over seismic performance because it is usually related to the robustness of a moment frame.Yet, this is not supported by the comparison in Fig. 12(b) which shows similar distributions of damage levels for buildings with 2 spans or less and those with more than 2 spans. Their average damage indices are also similar to each other.

3 FURTHER DISCUSSION 3.1 Terrain Effects

Since the inspected street goes along the mountain ridge, a majority of the inspected buildings were built on slopes. Several commonly-seen treatments for the foundation on a slope are summarized in Fig. 13(b)-(d). In the Type Ⅱ foundation, the slope is first filled with rubble to form a platform on which the foundation can be laid. The rubble fills are usually strengthened by concrete pillars. In the Type Ⅲ foundation, the foundation is laid at different elevations to fit the slope. Some rubble fill is necessary but the amount of fill is significantly smaller than that of Type Ⅱ foundations. In Type Ⅳ foundations, the slope is flattened for the foundation behind an RC retaining wall. If the building does not lean on the retaining wall, it is essentially the same as an ordinary building as in Type Ⅰ. However, most inspected buildings with the Type Ⅳ foundations leaned against the retaining wall. This makes them similar to those with stepped foundations in terms of dynamic characteristics. Along the in-plane direction of retaining wall, the existence of the retaining wall makes the stiffness of the bottom story much larger than the story above it, thus creating a stiffness discontinuity that may cause soft-story failure of the story above it. In the direction transverse to the slope, the stepped foundation and the retaining wall make the structural plan irregular so that detrimental torsional vibrations may be stimulated (Wang, 2010).

 Fig. 13 Typical foundations for RC frames in Charikot (a) Ordinary; (b) Foundation on rubble fill; (c) Stepped foundation; (d) Partial basement

Buildings C in Fig. 3 is a significant case that demonstrates the aforementioned problems associated with stepped foundations. From the street side, it appeared a typical weak story collapse at the bottom story of a 4-story building (Fig. 14(a)). From the backside, however, the collapsed story was actually the third story of a 6-story building if counting from the lowest foundation (Fig. 14(b)). The bottom columns on the street side were completely crushed, and as a result the whole building leaned towards the street.

 Fig. 14 Building C with stepped foundation (a) Street view; (b) Backside view

Although this specific example seems to suggest that buildings on slopes are more vulnerable to seismic damage or even collapse, the statistics of the building damage in the inspected region does not show higher vulnerability for the buildings on slopes. Among the 72 inspected buildings, eleven were ordinary buildings on flat foundations (Type Ⅰ), eight have Type Ⅱ foundations, fifteen have Type Ⅲ stepped foundation and twelve have Type Ⅳ foundations, whereas the foundation types of the remaining 26 buildings were unknown, including two that collapsed. Fig. 15 presents the damage levels of the inspected buildings by foundation type, excluding the 26 buildings of unknown foundation types. The average damage index, Da, for each group of buildings is also given in Fig. 15. The buildings with ordinary foundations (Type Ⅰ) did not exhibit less severe damage than other buildings in an average sense. The average damage index for all the buildings on slopes (that is Ⅱ + Ⅲ + Ⅳ) is 0.32, even smaller than that of buildings on a flat foundation (Type Ⅰ, Da=0.41).

 Fig. 15 Relationship between damage level of the inspected buildings and type of foundation

In addition, the geographic distribution of the building damage in the inspected region shows little correlation between the slope and the extent of the damage. As shown in the altitude graph of the inspected street in Fig. 16, all the collapsed buildings were concentrated in the lower half of the street. It is also obvious from Fig. 10 that the damage to the buildings in the lower half of the inspected street was much more severe than that in the upper half.

 Fig. 16 Altitude of inspected street with locations of completely collapsed buildings

Three sections of the mountain along the street are taken, namely, Section 1-1 and Section 2-2 where many buildings collapsed, and Section 3-3 near top of the mountain where most buildings were intact or only slightly damaged. The locations of these sections are given in Fig. 10. The cross sectional shapes of the three sections are depicted in Fig. 17, where the gradients of the slopes at the inspected street are also shown. The gradient of the slope in Section 1-1 is only slightly larger than that in Section 3-3, whereas the gradient in Section 2-2 is the smallest. This suggests that the magnitudes of the gradients are not correlated with the extent of building damage.

 Fig. 17 Altitude of three sections of the inspected region

Based on the above discussions, although the collapse and damage of several inspected buildings can be attributed to the structural problems associated with foundations on slopes, the slope is by no means the sole reason that explains the damage of the inspected RC frames in Charikot.

3.2 Pounding

Many instances of pounding between neighboring buildings were observed in the inspected area. For example, the partial collapse of Building E was a result of the pounding by Buildings D and F which fell upon it. Local damage was also observed on Building J that suggested severe pounding with Building K before both of them collapsed. Most collapsed buildings leaned against neighboring buildings and caused different extents of local damage to them.

In addition to the pounding associated with the collapsed buildings, pounding was also a source of damage to buildings that did not collapse. For example, the damage to the seven-story building shown in Fig. 18 was concentrated in the third floor where it pounded with a four-story RC frame on one side and a three-story RC frame on the other side. The damage was most clearly seen from the back side of the building and inside the building (Fig. 19). The damage included extensive collapse of the masonry infills and the concrete spalling of some RC columns.

 Fig. 18 Street side view of a seven-story building impacted on its neighboring buildings

 Fig. 19 Back side and interior views of a pounded building
3.3 Accumulative Damage

As shown in Fig. 2, Charikot is far from the epicenter of the main shock on April 25 but is very close to the April 26 MS7.1 aftershock and the May 12 MS7.5 aftershock. Locals confirmed that the inspected area was most affected by the April 26 aftershock. Many buildings were damaged and people were evacuated. There was further damage of some buildings in the May 12 aftershock. In extreme cases, the Buildings J and K (Fig. 20) were moderately damaged during the April 26 aftershock and completely collapsed in the May 12 aftershock. Fortunately, their collapse on May 12 did not lead to high fatalities (two passersby were killed) because they were already evacuated on April 26. This evidence suggests that the seismic damage to the buildings in Charikot may have accumulated during the multiple earthquake excitations.

 Fig. 20 Buildings J and K that were damaged on the April 25 main shock and collapsed on the May 12 aftershock

Such effect of accumulated damage was also reported in other places close to the epicenter of the May 12 aftershock. It was a 4-story RC frame in Chautara, the headquarters of Sindhupalchok district. It was heavily damaged during the April 25 mainshock, especially in the bottom story which exhibited significant residual drift (Fig. 21(a)) and extensive column hinging. During the May 12 aftershock, the bottom story completely collapsed (Fig. 21(b)).

 Fig. 21 A 4-story owner-built RC frame in Chautara (a) Before the May 12 aftershock; (b) After the May 12 aftershock (photos by owner of building)

Although the damage accumulation in some buildings was obvious during the Nepal earthquake sequence, it was difficult to quantify its effect by only field observations. Recent earthquake disasters have evoked studies on the effects of aftershocks and ground motion durations (Zhai et al. 2013, 2015, Hou et al. 2015). Yet, there are no explicit provisions on the seismic demand of accumulated damage particularly regarding the aftershock effect in seismic codes. Further study is needed on this issue.

4 CONCLUSIONS

The seismic damage to the owner-built RC frames in Charikot, the headquarters of Dolakha district, is summarised in this paper. By assessing the seismic damage to all buildings of the same structural system in a concentrated area, the seismic damage can be interpreted in a statistical manner so that the uncertainties and complexity associated with the damage characteristics and possible causes for such damage can be better addressed.

The local practice of RC frames in Charikot was close to the Nepalese ready-to-use guide for low-rise RC frames on an average sense. However, the owner-built RC frames easily exceeded the height limit for low-rise buildings in the guide, whereas the plan layouts and component dimensions and reinforcement remained similar. As shown by the damage statistics, the medium-rise buildings in the inspected area were much more fragile than the low-rise counterparts.

The damage characteristics of RC frames were also influenced by the rugged terrain of the inspected area. The collapse of several RC frames on the slopes can be attributed to their irregular stiffness distribution. On average, however, the extent of damage was not significantly different between buildings on slopes and those on flat foundations. Pounding between adjacent buildings and the effect of accumulative damage during the earthquake sequence were also significant.

REFERENCES
 AIJ. Report on the Hanshin-Awaji Earthquake Disaster—Building Series (Volume 1): Structural Damage to Reinforced Concrete Building [R]. Tokyo: Architectural Institute of Japan (AIJ), 1997 (in Japanese). AIJ. Preliminary Reconnaissance Report of the 2011 Tohoku-Chiho Taiheiyo-Oki Earthquake [R]. Tokyo: Architectural Institute of Japan (AIJ), 2012. 149-196. ASCE 7. Minimum Design Loads for Buildings and Other Structures [M]. American Society of Civil Engineers, 2010. BCJ. The Building Standard Law of Japan [S]. Tokyo: The Building Center of Japan, 2004. Chaulagain H., Rodrigues H., Jara J., Spacone, Varum H. Seismic response of current RC buildings in Nepal: A comparative analysis of different design/construction[J]. Engineering Structures, 2013, 49: 284–294. DOI:10.1016/j.engstruct.2012.10.036. GB 50011-2010. Code for Seismic Design of Buildings [S]. Beijing: General Administration of Quality Supervision, Inspection and Quarantine of the People's Republic of China, 2010 (in Chinese). GB/T 24335. Classification of Earthquake Damage to Buildings and Special Structures [S]. Beijing: General Administration of Quality Supervision, Inspection and Quarantine of the People's Republic of China, 2009 (in Chinese). Hicyilmaz K. Personal Communications, 2015. Hou H., Qu B.. Duration effect of spectrally matched ground motions on seismic demands of elastic perfectly plastic SDOFs[J]. Engineering Structures, 2015, 90: 48–60. DOI:10.1016/j.engstruct.2015.02.013. Humanitarian Data Exchange (HDX) (2015). 2011 Nepal census data (areas affected by the 2015 Earthquake). Downloadable at [https://data.hdx.rwlabs.org/dataset/2011-nepal], accessed on October 9, 2015. Japan International Cooperation Agency (JICA). The Study on Earthquake Disaster Mitigation in the Kathmandu Valley, Kingdom of Nepal, Volume Ⅰ: Summary [M]. Kathmandu: Ministry of Home Affairs, Her Majesty Government of Nepal, 2002a. 10-12, 77. Japan International Cooperation Agency (JICA). The study on earthquake disaster mitigation in the Kathmandu Valley, Kingdom of Nepal, Volume Ⅱ: Blueprint for Kathmandu Valley earthquake disaster mitigation. Kathmandu: Ministry of Home Affairs, Her Majesty Government of Nepal, 2002b. 156-177. Japan International Cooperation Agency (JICA). The study on earthquake disaster mitigation in the Kathmandu Valley, Kingdom of Nepal, Volume Ⅲ: Earthquake disaster assessment and database system. Kathmandu: Ministry of Home Affairs, Her Majesty Government of Nepal, 2002c. 29-43. Mahin S.A, Bertero V.V. Nonlinear Seismic Response of a Coupled Wall System[G]. Proceedings of the ASCE National Convention, 1975. NBC 205. Ready to Use Guideline for Detailings of Low Rise Reinforced Concrete Buidings Without Masonry Infill [M]. Ministry of Urban Development, Government of Nepal, 2012. Shakya K., Pant D.R., Maharjan M., Bhagat S., Wijeyewickrema A.C., Maskey P.N. Lessons learned from performance of buildings during the September 18, 2011 earthquake in Nepal[J]. Asian Journal of Civil Engineering, 2013, 14(5): 719–733. Sun B.T, Yan P.L. Damage characteristics and seismic capacity of buildings during Nepal MS8.1 earthquake[J]. Earthquake Engineering and Engineering Vibration, 2015, 14(3): 571–578. DOI:10.1007/s11803-015-0046-x. Deger Z.T, Wallace J.W. Collapse assessment of the Alto Rio building in the 2010 Chile earthquake[J]. Earthquake Spectra, 2015, 31(3): 1397–1425. DOI:10.1193/060812EQS209M. United Nations High Commissioner for Refugees (UNHCR) (2015). Nepal: 2015 earthquakes and aftershocks, People killed/injured by district. Downloadable at [http://reliefweb.int/map/nepal/nepal-2015-earthquakes-and-aftershocks-people-killed-injured-district-03-jun-2015], accessed on Oct 9, 2015. United Nations Office for the Coordination of Humanitarian Affairs (UNOCHA) (2015). Nepal Earthquake: Weekly situation update, 7 August 2015. Nepal Earthquake Assessment Unit, UN. Villaverde R. Explanation for the numerous upper floor collapses during the 1985 Mexico City earthquake[J]. Earthquake Engineering & Structural Dynamics, 1991, 20(3): 223–241. Wang L.P. Design Ground Motion Input and Control Method of Lateral Stiffness for Building Structures on the Slope [D]. Ph.D. Thesis. Chongqing: Chongqing University, 2010 (in Chinese). Ye L.P, Qu Z., Ma Q.L, Lin X.C., Lu X.Z., Pan P. Study on ensuring the strong column-weak beam mechanism for RC frames based on the damage analysis in Wenchuan earthquake[J]. Building Structures, 2008, 38(11): 52–59 (in Chinese with English abstract). Zhai C.H., Wen W.P., Chen Z.Q., Li S., Xie L.L. Damage spectra for the mainshock-aftershock sequence type ground motions[J]. Soil Dynamics and Earthquake Engineering, 2013, 45: 1–12. Zhai C.H., Wen W.P., Li S., Xie L.L. The ductility-based strength redution factor for the mainshock-aftershock sequence-type ground motions[J]. Bulletin of Earthquake Engineering, 2015, 13: 2893–2914. DOI:10.1007/s10518-015-9744-z.