NEWSLETTER

Vol.8 No.2 NOVEMBER 2004 ISSN:1174-3646

Earthquake Hazard Centre Newsletter, Vol.8 No.2, November 2004

Supported by Robinson Seismic Ltd

Contents

Editorial p.1Principles of Earthquake Resistance p.2

p.3

p.5

p.7

Earthquake Resistant Earthen Buildings

Effectiveness of the Tropical Cyclone Warning Systems in Tikopia and Anuta

Performance of School Buildings in Bingol during the 1st May 2003 Earthquake The Environmental and Societal Impacts of Cyclone Zoe and the

www.robinsonseismic.com

1

EditorialHurricanes and earthquakes

Last month, during mid September, the media focused considerable attention on Hurricane Ivan. Reports traced its destructive path across the Caribbean as it gutted homes, killing over 50 people and inflicting billions of dollars in damage. In Grenada, over ninety percent of the buildings were damaged or destroyed. Some of the images of the damage Hurricane Ivan caused could well have been taken immediately after a severe earthquake. It was clear that buildings had been subjected to immensely strong lateral loads. One of the questions that this damage raised is this: is it feasible to design against winds with speeds up to 260 km/h, or 72 m/s? Well, the answer has to be a conditional yes. Technically it is feasible, but it is also expensive and requires high quality engineering design and construction.

The basic design wind speed in the New Zealand loadings standard is 50 m/s. Buildings are usually designed for lower speeds after taking account of terrain, height above ground and other factors. So for a NZ building to resist hurricane-like winds, loads would be increased approximately by a factor of two. Larger and greater numbers of structural members at closer centres would be required, as well as considerably stronger fixings and fasteners. Claddings also need to be stronger and require more frequent support from members like studs, girts, rafters and purlins. Such additional costs and design rigour might only be justified for certain buildings, but as we have seen, the consequences of hurricane damage on construction below this standard can verge on the catastrophic. Such high wind loads would mean wind

loading would often be a more severe lateral load case than earthquake loads.

So what are the similarities and differences between the loadings imposed by each of these natural disasters?

Both loads act laterally, but hurricanes also generate enormous uplift pressures on roofs. Roofs, of all surfaces on a building, experience the greatest pressures. Frequently roofs are torn off, and once they go so do the roof diaphragms that are needed to distribute the lateral loads to bracing elements parallel to the wind direction. The severity of the damage is then immediately intensified as face-loaded walls find themselves with no support at roof level and are therefore prone to collapse.

Regarding dynamic effects, hurricane damage is unaffected by the dynamic characteristics of buildings. Wind loading is far more sustained than that caused by earthquakes, with the duration of severe gusts lasting in the order of three seconds or longer. The lateral load resisting structure is therefore fully tested and must be designed for the full magnitude of the load. Unfortunately, unlike earthquake loads, neither actual wind loads nor design wind loads are reduced by structural ductility.

To resist both sets of loads successfully, buildings must possess horizontal and vertical lateral load resisting elements. Horizontal elements include roof, ceiling or floor diaphragms. These are the first line of defence. They transfer the loads to vertical elements like moment-resisting frames, cross-braced frames or shear walls that act parallel to the loads, and in turn provide a load path to the foundations. Hurricane resistance requires one additional feature strong vertical tie-downs of roof structural elements.

Hurricane resistance also benefits from the mass of construction. The heavier the roof and walls, the less likelihood there is of the roof and the superstructure being lifted from the foundations. Unfortunately increased mass means increased seismic inertial loads. So countries like the Pacific States that can expect both of these natural hazards to strike need to be aware of the relative risks of occurrence of each when choosing their construction materials.

Earthquake Hazard Centre Newsletter, Vol.8 No.2, November 2004

�

shear wall resists all X-direction loads

Ye

YX

Sshear wall resists y-direction loads

Sshear wall resists x-direction loads

Ggravity-only column

� lev r armeL

diaphragm above twists about the CoR

CoM

CoR (Centre of Resistance)

As the gravity-only columns are too flexible and weak to resist lateral loads there is no vertical structure to resist the torsional moment Ve. The building is torsionally unstable. A solution is to make one of the walls “non-structural” and replace its lateral load capacity by say moment-resisting frames acting in the same direction and widely separated to achieve large lever arms between them. Not only does this design eliminate torsion when earthquake loads act in the Y-direction, but the long lever arm between the frames provides a torsional stabilizing couple when loads act in the X-direction as shown below.

Principles of Building Earthquake Resistance No. 30

Building torsion - how to resist it (2)

�

Plan of a torsionally eccentric building

CoM(Centre of Mass)

YY

YY

YX

�V

V

new frame with a non-structural infill wall separated from the frame

new moment-resisting frame resists half the loads in the -direction. The other frame also resists half the load.

ddiaphragm twists about CoR

Force in the frame generated by the horizontal displacement in the Y-direction caused by the twisting diaphragm when multiplied by the lever arm equals Ve, so the building is torsionally stable.

Ye

V

V

earthquake load acts at CoMearthquake load is resisted by shear wall because it is by far the strongest element in the Y-direction

CoR

2

Earthquake Hazard Centre Newsletter, Vol.8 No.2, November 2004

Summary of “Earthquake Resistant Earthen Buildings” by Marcial Blondet

thand Gladys Villa Garcia, 13 World Conference on Earthquake engineering , Vancouver, B.C., Canada, August 1-6, 2004.

Seismic Behaviour of Earthen BuildingsMany countries that traditionally use soil as a building material are located in areas of high seismic hazard. Unfortunately, the seismic performance of traditional unreinforced earthen buildings is extremely poor, as explicitly demonstrated during the recent Bam and Morocco earthquakes.

As a construction material, soil is heavy, brittle and weak. During seismic motion, heavy walls develop large inertia forces, which they are unable to resist. Thus, they suddenly fracture and many times collapse, preventing evacuation and causing death and material loss.

In Peru, several historic earthen constructions have been able to withstand severe earthquakes because of their massiveness and regular configuration. Nevertheless, the latest earthquake in Iran (December 2003, M 6.6), which w

destroyed thousands of poorly made adobe houses and damaged important ancient historical monuments such as the earthen citadel of Bam raised questions concerning the argument for massiveness as a guarantee for earthquake endurance.

“Modern” earthen dwellings in many countries are built to imitate the architectural features of clay brick masonry houses. Thus, earthen houses are built in a similar fashion without any structural reinforcement, with several storeys, thin walls, large window and door openings and irregular plan and elevation configurations (Figs.1 and 2). These buildings are extremely vulnerable to earthquakes. When an earthquake occurs, out-of-plane seismic forces produce large vertical cracks at the corners, and in-plane shear forces produce diagonal cracks in the walls. The walls are thus broken into large independent pieces, which fall down, causing in turn the collapse of the roof.

The M 7.8 Huaraz earthquake of 1970 caused the death w

of around 70 000 people, about half of them buried under their adobe houses. This tragedy raised the consciousness in Peru about the need to investigate the problem of seismic strength of adobe buildings. Initially, the research at the Catholic University of Peru (PUCP) was oriented towards the experimental study of several different alternatives of structural reinforcement using rural materials. This was done with 4x4 meter reinforced concrete tilting platform used to test full-scale adobe modules (Figs.3 and 4). The seismic force was represented by the lateral component of the weight of the modules. The main conclusion was that an interior reinforcement made of vertical cane, combined with an arrangement of horizontal crushed cane every fourth layer of adobe blocks substantially increased the seismic strength of the adobe modules.Figs.1 and 2 “Modern” adobe houses in Peru.

Figs.3 and 4 Full-scale adobe modules over tilting platform.

3

Earthquake Hazard Centre Newsletter, Vol.8 No.2, November 2004

An important research project was started in 1992 where the seismic simulator of the structures laboratory of the PUCP was used to perform dynamic testing of full-scale adobe modules. The main conclusions of this project were that improvement in the construction technique (quality of labour and materials) by itself increased the strength and stiffness of uncracked walls, but had negligible influence after significant cracking occurs. During severe shaking, only the presence of horizontal and vertical cane reinforcement, combined with a solid collar beam, were able to prevent the separation of the walls in the corners, thus maintaining the integrity of the structure. In consequence, this reinforcement is effective in preventing or delaying the collapse of the building.

Following that, an experimental project was funded in 1996 to develop simple techniques to reinforce existing adobe buildings. “U” shaped walls were tested on the seismic simulator with different reinforcement materials like wooden boards, rope, chicken wire mesh and welded mesh. The best results were obtained with welded mesh (1mm wire spaced at 20mm), nailed with metallic bottle caps against the adobe and covered with cement-sand mortar (Fig.5). The mesh was placed in horizontal and vertical strips, simulating beams and columns. After successful testing of four full-scale modules on the seismic simulator, this solution was applied to the reinforcement of existing adobe houses located in different regions of Peru (Fig.6). In 2001, a strong earthquake (Magnitude 8.4) occurred in Southern Peru and destroyed most adobe houses in the affected region. However the houses that were reinforced with welded mesh located in the region suffered no damage and were used as shelters.

Adequate Reinforcement TechniquesThe reinforcement systems have their limitations. The main limitation of cane is the fact that it is not available in all regions. Moreover, even in areas where cane is produced, it is practically impossible to obtain the

required quality for a massive construction or reconstruction program. External reinforcement with welded mesh cost around US$200 for a typical one floor, two room adobe house. This amount exceeds the economic capacity of the Peruvian adobe user, whose monthly income is most likely the legal minimum of less than US$150.

Therefore, the proposed reinforcement systems, technically efficient for earthquake endurance, are still far from being real alternatives to improve the seismic behaviour of adobe houses. It seems imperative to continue research to develop reinforcement systems that use industrially produced materials, acceptable to adobe users because of their low cost and simplicity of application.

Recent and Ongoing ResearchA preliminary research project is currently being developed with the objective of studying the technical and economical feasibility of using locally available industrial products such as wire, plastics and geo-synthetics in the seismic reinforcement of adobe houses. Six ‘I’ shaped full-scale adobe walls, with and without reinforcement, are being constructed at the Structures Laboratory and will be tested under horizontal cyclic loading. The geometric configuration, including a small central window, and the overall dimensions are similar for all the walls. Three baseline walls include one traditional unreinforced, one with interior cane reinforcement and one with welded mesh placed externally stimulating beams and columns. The reinforcement proposals include:

1. Adding ½” diameter PVC water pipes as inner vertical reinforcement, combined with a low-cost ½” plastic mesh embedded horizontally in the mortar every fourth layer. Both reinforcement elements are tied together with plastic string. The PVC pipes are anchored in the foundation beam and in the upper tie beam, both made of concrete.2. Placing vertical ½” diameter corrugated steel bars

inside the walls at the corners, with no horizontal reinforcement.3. Fixing a geo-synthetic mesh externally to the

walls, simulating columns and beams. The covering is made of mud mixed with straw.Fig.5 Metallic bottle caps and steel mesh nailed against

adobe wall.

Fig.6 Typical reinforced house in Peru.

4

Earthquake Hazard Centre Newsletter, Vol.8 No.2, November 2004

The total column area of buildings with moment-resisting frames was approximately 1% of the floor area regardless of the number of floors. Consequently, the number of floors was the determinant for the seismic performance of the structures. The level of damage of the lateral load resisting system relative to the number of floors can be categorised as follows:

· 5 two-storey schools: 4 moderately damaged, 1 lightly damaged

· 11 three-storey schools: 3 collapsed, 6 severely damaged, 2 moderately damaged

· 1 four-storey school: 1 severely damaged

The three and four-storey buildings typically sustained severe masonry wall damage. There were several construction and structural design deficiencies commonly observed in the school buildings. In most of the structures surveyed, the quality of construction practices was uniform. Specific problems noted were:

· use of unwashed aggregate,

· use of aggregates with large maximum size (up to 100 mm),

· use of undeformed bars,

· inadequate preparation of cold joints

One of the most common structural problems observed in these buildings was the presence of captive columns, which made the structures vulnerable with respect to column shear failures. In almost all the schools, openings for small windows in the furnace room and restrooms were placed adjacent to columns. The exterior rectangular columns were oriented with the strong axis resisting

ConclusionExperimental research has shown that it is possible to mitigate the effects of strong earthquakes on earthen buildings, delaying or even preventing their collapse. However, the scope of the problem is much bigger, because it involves acceptance by users, availability of adequate materials and further economical aspects. The challenge is to continue exploring new alternatives until a technically good solution is found that will contribute to solving the housing deficit in most impoverished third world countries and at the same time guarantee survival during strong earthquakes.

For a copy of the full paper contact the EHC.

Summary of the paper “Performance of School Buildings in Bingöl during the 1st May 2003 Earthquake” by T. Gur, J.A. Ramirez, M.A. Sozen et al.

IntroductionAn earthquake of Magnitude 6.4 (M ) struck the city of w

Bingöl in eastern Turkey on 1 May 2003, resulting in 168 deaths and extensive damage to private and public buildings. This paper focuses on the damage to 23 reinforced concrete schools and 4 dormitories. The sample includes 2, 3 and 4-storey moment-frame structures, some with only masonry infill walls and others with both masonry infill walls and reinforced concrete shear walls.

The number of storeys, type of the structural system, ratio of reinforced concrete column and wall areas to the overall floor area, and proximity of the schools to the epicenter were selected to be the major study parameters. Regardless of the school type, total collapse was prevented if reinforced concrete shear walls were present because they compensated for the vulnerability of the captive columns effectively shortened by the masonry walls. The salient feature of the structural damage in Bingöl was defined by the varying interaction, ranging from good to very bad, of the tile masonry with reinforced concrete frames.

School Buildings with a RC Moment-Resisting Frame System Of the 17 buildings in this category, 16 had the same column layout. The lateral load resisting system in these buildings were regular in plan and the majority of the columns were aligned in regular bays with most of the beams framed into the columns. The dimensions of the columns in the buildings were typically 0.3m x 0.5m and their orientation were the same in all but one building. The locations of the masonry infill walls varied depending on the use of the space in the schools. Furthermore, the exterior masonry walls were typically thicker than the interior walls.

Fig.7 and 8 Shear failure of captive columns created by the small windows.

5

Earthquake Hazard Centre Newsletter, Vol.8 No.2, November 2004

moments in the short direction of the building layout. Therefore, windows on the exterior walls in the long direction of the building exposed columns to shear forces acting perpendicular to their weak axis for bending (Figs.7 and 8). It was observed also that the crushing of the masonry walls in the upper corners created captive columns (Figs.9 and 10).

In the school buildings that were visited, the detailing of the structural members was inadequate with regards to requirements of modern seismic codes. Lack of confinement in plastic hinge regions of the columns was observed to be one of the most significant causes of damage. Even though the spacing of stirrups was reduced in the end regions of some columns, the amount of transverse reinforcement provided was not sufficient to prevent shear failures, particularly in the case of captive columns. Another detailing deficiency commonly observed was the inadequate anchorage of the free ends of the stirrup reinforcement and the lack of face load support to masonry infill walls (Fig.11 and 12).

Comparison of the Performance of the Frame and Dual SystemsBuilding damage observations indicate that the

performance of dual systems was satisfactory. Even though some of the dual system buildings were rated as severely damaged because of the damage associated with captive columns and cold joints, observed damage to the masonry walls indicate that the reinforced concrete walls were effective in controlling lateral drifts. Buildings with moment resisting frame systems did not perform well during the earthquake. Although the quality of construction is quite uniform for all the buildings, frame systems were more vulnerable to damage associated with deficiencies in construction practice. The flexibility of moment frame buildings resulted in larger drift demands than those in buildings with dual systems, which cause severe damage and in many cases the collapse of the structure due to loss in gravity load capacity. Further, the damage level of infill masonry walls in moment frame buildings that were severely damaged supports the conclusion regarding the role of structural walls in controlling drift demands. The shear damage to columns was very severe in buildings with moment resisting frames.

ConclusionsThe observations in the school buildings showed that structural walls improve the behaviour of reinforced concrete systems drastically. Structural walls are especially recommended in the structural design of school buildings using reinforced concrete frames such as those found in Bingöl.

For a copy of the full paper contact the EHC.

Figs.9 and 10 Shear failure of captive columns caused by crushing at the upper corner of masonry walls.

Figs.11 and 12 Collapse of free-standing masonry walls onto the beds in dormitory buildings.

6

Earthquake Hazard Centre Newsletter, Vol.8 No.2, November 2004

A s u m m a r y o f t h e p a p e r “ T h e Environmental and Societal Impacts of Cyclone Zoe and the Effectiveness of the Tropical Cyclone Warning Systems in Tikopia and Anuta” by L. Anderson-Berry (Bureau of Meteorology Australia / JCU Centre for Disaster Studies), Chanel Iroi and Alan Rangi (Solomon Island Weather Services).

IntroductionSevere Category 5 Tropical Cyclone Zoe was the most intense cyclone ever to be forecast and monitored by the Fiji Meteorological Service in the southwest Pacific region. Weather services around the world watched Zoe's development and movement with interest as the system intensified and concern was raised that landfall may eventually be made in Fiji.

International interest in the progress and impact in this system was intensified when it was evident that Tikopia and Anuta in the Solomon Islands, two of the world's smallest and most remote islands, had suffered the full fury of the storm over several days. Communities living on these islands were known to have been without two-way communication with national and international weather services and the possibility that they had been unwarned and unprepared was very real.

TikopiaCyclone Zoe devastated Tikopia. High winds began battering the island late in the evening of Friday December 27, 2002, and for the next three days the islands and its village communities were relentlessly pounded with cyclonic winds, storm surge and wind-driven waves. When the storm finally subsided, the full extent of Zoe's destructive forces began to be realised Initial delays to an early emergency response resulted in the medical assessment team not arriving on Tikopia until Sunday January 5, 2003.

Structural DamageStructural damage was most severe in Ravenga on the eastern side of the island that suffered the effects of rain, destructive winds and storm surge. Faea on the western side was exposed to rain and severe winds (Figs.12 and 13). Overall it was estimated that approximately 70% of the total village housing was severely damaged or destroyed.

Most of the remaining 30% of dwellings were somewhat damaged. A total of 176 huts in Ravenga, almost all the dwellings in the district, were totally destroyed, many were completely washed away. Numerous residents lost all their belongings and escaped with only the clothes they were wearing. These families are now living in very rudimentary shelters built from debris and black plastic sheeting. A majority of the 153 houses in Faea were also damaged or destroyed. Some with limited damage have been repaired with broken sago palm leaves. These

repairs are a good temporary measure but are unlikely to last in the longer term. Approximately 330 dwellings across the islands are in need of replacement or substantial repair. In addition to household dwellings most community buildings, constructed with both traditional materials and sawn timber with corrugated iron roofs were washed away, one was destroyed and three sustained moderate damage. Both of the primary schools were destroyed with all educational material lost (Fig.14). The clinic building was damaged but was still usable by the medical assessment team. Substantial repairs are urgently required. One of the two trades stores in Ravenga village was totally destroyed, the other building was badly damaged but the stock was saved. The vast majority of buildings on Tikopia are traditional structures constructed with local materials. The normal life span of sago palm roof thatch is just two years; therefore all surviving structures on the island will need new roofs within the next 12-18 months. All traditional housing material on the island has been lost. It will take 6-12 years for sago palms

Figs.12 and 13 Damaged huts in Ravenga and Faea.

Fig.14 Remains of one of the schools.

7

Earthquake Hazard Centre Newsletter, Vol.8 No.2, November 2004

to regenerate and even longer for structural timber to be milled. Tikopia will therefore not be able to approach self-sufficiency in building materials for at least 12 years.

AnutaThe impact of Cyclone Zoe was less devastating here than it had been on Tikopia; the island was nevertheless battered for almost 3 days resulting in significant wind and storm surge damage.

Structural DamageCyclone damage to built structures in Anuta was less than that experienced in Tikopia. The majority of buildings are constructed with bush materials: sago palm thatched roofs and walls covering bamboo and timber frames. The main fastenings are twine made of natural vines or nylon and nails. Five dwellings were destroyed during the cyclone, a further 24 were badly damaged, and several more were somewhat damaged. Much of the minor damage has been repaired using broken sago leaves but this is not expected to last long.

Most residents were able to secure their possessions. Roofs were often protected with palm fronds and banana trunks that were cut down and laid against the roof and walls. Severe damage from the storm was mitigated with the three metre high sea wall that protects the main village along the east coast. The wall is constructed of unmortared coral blocks and is well maintained by the people whose homes are built behind it. The walls withstood the waves well and were damaged only where trees fell on them. One of the two churches lost its roof, but this had already been rethatched when the assessment team arrived. The Tematai primary school, constructed of sawn timber corrugated iron was not damaged. However, it is in extremely bad repair and it is poorly resourced. With students about to return to school, school-packs sent by World Vision will help the problems, but more resources are urgently needed.

The severe damage to the natural environment means that no sago palms for repairs or rebuilding are available locally. This will remain the case for many years.

Warning EffectivenessPeople that receive timely warning messages generally began preparations immediately. This was limited to cutting palm fronds and banana trunks and laying them on roofs to support and strengthen roofs and walls and confining chickens to their coups.

The design of the village huts with low walls and sloping thatched roofs may be considered to be a long-term cyclone mitigation strategy as this design typically withstands cyclones although many that were located on the near-shore beach areas were exposed and unprotected from both wind and surge effects. In Anuta, the 3 metre sea-wall along the eastern edge of the beach just behind the high tide line that has been constructed from coral

block and is well maintained and provided an effective protective barrier. During Cyclone Zoe, waves over-topped this wall and sand and salt water spread through the village wetting property and spoiling crops. Dwellings however were saved as they were protected from the force of the powerful storm surge.

Tikopian and Anutan communities were not well warned for Cyclone Zoe and neither was particularly well prepared. They have some experience of living with cyclones, although it is more than 50 years since a cyclone of Zoe's intensity had been experienced. Even then as one elderly villager pointed out, “that cyclone lasted only three hours; this one lasted three days!” Reponse to warnings was appropriate albeit a little bit late. Damage to both island communities was severe and losses are catastrophic. The fact that no lives were lost may possibly be attributed to a combination of many factors including, the communities' understanding and knowledge of their landscape, well developed intuitive self-protective behaviours, the supportive societal networks that foster a sense of mutual community responsibility, a resilience arising from communities relative isolation and perhaps also a lot of 'good luck'. It must be clearly stated and always remembered that, in terms of loss of life, the actual outcome of this event was absolutely the most unlikely outcome.

A full version of this article can be found at the James Cook University website:

(Acknowledgements: Loti Yates and his team at The Solomon Islands National Disaster Council; Members of National Disaster Council Central Control Group; NDC / OCHA Assessment Team: Rex Tara, Nancy Jolo, Ambrose Kirei, Dominic Tua, Ian Aujare, Alfren Inomae, Herman Oberli; Charlie Higgins OCHA; Mr Bob Davis Australian High Commissioner in Honiara; Stacy Greene, Second Secretary, Australian High Commission Honiara; Johnson Honomae SIBC; Geoff Miller, AusAid; Steve Banks EMA.)

www.tesag.jcu.edu.au/CDS/index.htm

Earthquake Hazard CentrePromoting Earthquake-Resistant Construction in Developing CountriesThe Centre is a non-profit organisation based at the School of Architecture, Victoria University of Wellington, New Zealand. It is supported financially by Robinson Seismic Ltd and the Ministry of Civil Defence and Emergency Management.Director ( honorary) and Editor: Andrew Charleson, ME.(Civil)(Dist), MIPENZ; Research Assistant: Kirsten LimMail: Earthquake Hazard Centre, School of Architecture, PO Box 600, Wellington, New Zealand.Location: 139 Vivian Street, Wellington.Phone +64-4-463 6200 Fax +64-4-463 6204E-mail: quake@arch.vuw.ac.nzThe Earthquake Hazard Centre Webpage is at :-http://www.ehc.arch.vuw.ac.nz

8