Construction and design defects in the residential buildings and observed earthquake damage types in Turkey

Turkey is situated in a very active earthquake region. In the last century, several earthquakes resulted in thousands of deaths and enormous economic losses. In 1999, the Kocaeli earthquake had an approximate death toll of more than 20 000, and in 2011 the Van earthquake killed 604 people. In general, Turkish residential buildings have reinforced concrete structural systems. These reinforced concrete structures have several deficiencies, such as low concrete quality, non-seismic steel detailing and inappropriate structural systems including several architectural irregularities. In this study, the general characteristics of Turkish building stock and the deficiencies observed in structural systems are explained, and illustrative figures are given with reference to the Turkish Earthquake Code 2007. The poor concrete quality, lack of lateral or transverse reinforcement in beam– column joints and column confinement zones, high stirrup spacings, under-reinforced columns and over-reinforced beams are the primary causes of failures. Other deficiencies include weak-column–stronger-beam formations, insufficient seismic joint separations, soft-story or weak-story irregularities and short columns. Similar construction and design mistakes are also observed in other countries situated on active earthquake belts. Existing buildings still have these undesirable characteristics, and so to prepare for future earthquakes they must be rehabilitated.


Introduction
In Turkey, 70 % of the population is living in the first-and second-degree seismic zones, and 95 % of the buildings are at risk.Losses have been experienced from medium-intensity and severe earthquakes for many years.The losses were not limited to rural regions; there were also major financial losses and other intangible damages in urban regions (Erzincan 1939and 1992, Kocaeli 1999, Van 2011, etc.).In those urban regions, most of the structural stock was built from reinforced concrete.Medium-intensity earthquakes in particular can be endured as an ordinary event in developed countries, but they are still considered one of the most important natural disasters in Turkey.Because most of the current buildings in Turkey were constructed before Turkish Earthquake Code 2007 (TEC 2007), their earthquake-resistance features are insufficient and their structural irregularities pose a danger.Another important point is that many of the buildings that have structural irregularities are high-rise buildings.
Large earthquakes occur in various regions of Turkey in 10-15 year intervals.Earthquake codes are also overhauled at certain times.Earthquake codes are generally reconsidered in response to application and design mistakes observed in previous earthquakes and in response to the failure mechanisms observed in damaged buildings.In the practice of structural engineering, it is necessary to determine and classify the mistakes and to prevent them from occurring in future applications.Thus, it is important to study the damages caused by earthquakes and to determine how they occur.It is difficult to understand the damage mechanisms of wholly collapsed buildings after earthquakes.For this reason, it is better to concentrate on moderately and heavily damaged buildings in the technical investigations and damage assessments that are conducted after earthquakes.As a result of investigations in these buildings, the engineering and application mistakes can be taken into consideration.The collapse of an entire building or the loss of a floor generally occurs as a result of similar mistakes.In this study, the most frequent mistakes during the design, construction and usage stages in  reinforced concrete buildings and their results after earthquakes are described via visual elements.The observations and the reasons for mistakes are evaluated with consideration of the TEC 2007.The main principles of earthquake-resistant structural design are presented in the following codes: -In light earthquakes, structural and nonstructural system elements in buildings should not be damaged.
-In moderate earthquakes, the damage to structural and nonstructural system elements should be at a level that can be repaired.
-In heavy earthquakes, the partial or whole collapse of buildings should be prevented in order to avoid loss of life (Onen et al., 2011).
In Turkey, it is not uncommon for materials to be stolen or for unqualified materials to be used in construction.When damages, collapses and losses occur after earthquakes due to these activities, it is usually just the building contractors who are convicted.However, the problem is not that simple.The causes of damage in affected buildings (1992Erzincan, 1994Dinar, 1998Adana-Ceyhan, 1999Gölcük, 1999 Düzce and 2011 Van) can be ordered as follows: -load-bearing system mistake (strong beam-weak column, inadequate size, inadequate bearings or misplacement of bearing elements, etc.) -architectural design mistake (ribbon window, soft story, side discontinuities, incorrect geometrical structuring, etc.) -low-quality materials (low concrete strength, poor reinforcement, etc.) -floor effects.

Earthquakes in Turkey
The earthquakes in Turkey, a country on the Alpine-Himalayan seismic belt, are related to the movement of the African and Arabian plates in a north-northeast direction.
The movement depends on the spreading of the Atlantic into the Red Sea.These plates dive under the Eurasian plate causing compression in the East Anatolian Region between the Arabian plate, and the Eurasian Continent.This compression sets large faults into motion, including the North Anatolian fault and the East Anatolian fault.
The North Anatolian fault is 1400-1500 km long.The Anatolian plate between the North Anatolian and East Anatolian faults moves west at 13-27 mm per year and proceeds toward Crete by curling to the left in the west.The plate motions in and around Turkey are shown in Fig. 1.

Seismic zones in Turkey
Turkey is in the Alpine-Himalayan region, which is one of the most active seismic zones in the world.Approximately 42 % of the country's surface area is in the first-degree seismic zone.There are five seismic zones in Turkey: 1. First-degree seismic zone: the areas close to active faults; earthquakes occurring here cause substantial loss of life and property.
2. Second-degree seismic zone: these are the areas in which earthquakes cause less damage than in the firstdegree seismic zone.
3. Third-degree seismic zone: shocks cause less damage in these areas.
4. Fourth-degree seismic zone: shocks cause little or no damage in these areas.
5. Fifth-degree seismic zone: there are few shocks in these areas, or no shocks are felt (Fig. 2).
3 Overview of turkish structural stock and earthquake damages

Overview of structural stock in Turkey
Research revealed that 70 % of the building stock in big cities such as Istanbul, Ankara and Izmir is illegal and that 50 % of the building stock in the country has no legal construction authorization.
The Ministry of the Environment and Urban Planning stated that "half of the buildings constructed before 2002 are non-resistant, and for this reason there are a great number of buildings that need seismic performance improvement and reinforcement" (Pampal, 2000).
The biggest city in Turkey is Istanbul and it is located in first-and second-degree seismic zones.According to inventory studies by the Ministry of Environment and Urban Planning in 2008, there are approximately 1 million buildings in Istanbul alone, and no engineering services were found for 400 000 of them.Engineering services in others were found to be inadequate for mitigating disaster risks.First, in order to mitigate risk, it is necessary to identify buildings with problems and to improve them.However, Istanbul has a building stock of 1 million.One-by-one reinforcement of buildings will not be sufficient to eliminate urban risks in building areas exposed to danger (Kizilkanat et al., 2011).
The third biggest city, Izmir, is also located in first-degree seismic zone and the situation of reinforced concrete structures is the same as Istanbul.At the Izmir Symposium on Reducing Disaster Risk (2009), it was determined that the building quality in Izmir was 3 % good, 52 % medium and 45 % poor/bad according to studies carried out in three pilot regions (1490 buildings) by the Izmir Chamber of Civil Engineers.This study was to serve as a model for structural stock in Izmir.In the study, it was found that 60 000 (8.2 %) of the 725 000 buildings in Izmir will endure heavy damage and that approximately 70 000 (9.5 %) will endure moderate damage in the case of an earthquake.
The situation in central and eastern Turkey is similar to Istanbul and Izmir.On Sunday, 23 October 2011 at 13:41 local time (LT), destructive ground motion occurred (M w : 7.2) with an epicenter at Tabanli, a village in the Ercis district of Van city (AFAD, 2011).Destruction and damage to buildings occurred in Van and other Ercis district centers and villages, and losses of life and property were experienced because of the building damage.The Van-Ercis earthquake caused severe destruction in the center of Ercis district due to the building stock of the region, but similar levels of destruction were not observed in the center of Van (Onen et al., 2011).Following the earthquake on 23 October, another earthquake (M w : 5.6) occurred on Wednesday, 9 November 2011 at 21:23 LT, with an epicenter in the Edremit district of Van city.In total, 644 citizens lost their lives in the earthquakes on 23 October and 9 November; 1966 citizens were injured and 252 citizens survived the wreckage.In the first week of the earthquake, 187 earthquakes with magnitudes of 4.0-4.9 and 13 earthquakes with magnitudes greater than 5 occurred in the region, and approximately 180 aftershocks occurred every day for the first month after the earthquake (AFAD, 2011).Every earthquake is unique in terms of the characteristics of the ground motions and the responses of buildings to those ground motions.The calculations of civil engineers include many assumptions related to the design of the structures.First, the concrete material used is not homogeneous, and the properties change over time.Cracking occurs in reinforced concrete elements, which complicates the engineering assumptions related to concrete.In addition to these uncertainties about materials, the determination of dead and live loads for building has its own uncertainties and assumptions.When building design is discussed only in terms of vertical load, the work of a civil engineer is relatively easy.However, the possibility of earthquake loads must be considered, as many uncertainties are introduced.The most effective method to prove the validity of an engineering problem that includes such uncertainties and assumptions is to perform experiments.However, doing full-scale three-dimensional experiments with buildings is expensive and difficult.In experimental studies, experiments are generally performed on samples that are modeled, usually in two dimensions, by minimizing with a scale factor.From an engineering point of view, it is possible just a few years after the completion of a building to understand whether the building is resistant against vertical loads.In this way, an assessment can be made about the accuracy of the assumptions made in the design of the building.However, for earthquake calculations, it is necessary to wait for the occurrence of an earthquake before the assumptions about earthquake-resistant structure design and the calculations of earthquake loads can actually be tested and validated.For this reason, it is important to examine the causes of damage and collapse after an earthquake in order to revise the earthquake code, review current engineering assumptions, reorganize the code for building construction and, above all, raise awareness among civil engineers and architects.

Short column behavior
The most important defects observed in damaged structures after earthquakes are mistakes related to deficient architecture or load-bearing systems.It is very difficult and sometimes impossible for a civil engineer to convert an existing structural system to an earthquake-resistant one.The first category of architectural design mistakes commonly encoun-tered after earthquakes is the formation of short column behavior.
If there is a difference between the lengths of columns on the same floor, it can become a damage mechanism during an earthquake (Fig. 3).Ribbon windows that are made from column to column in the halls of structures such as schools and hospitals and the presence of which are not considered in static projects can cause brittle fractures in columns.
Short columns form in structures mainly because of brick walls that are not load bearing.Brick walls (partition walls) are considered to be dead loads in structural analysis.However, brick walls that are accepted as non-load-bearing contribute to load transfer in the frame under horizontal loads as diagonal struts.They increase the rigidity of the frame and limit horizontal displacement.In some situations, ribbon windows can be made in the upper parts of brick walls just under the beam.This situation is frequently encountered in the illumination of halls in dormitories and training centers as well as in basements.However, partial brick walls extending from column to column will prevent the horizontal translation of columns and cause the formation of short columns (Fig. 4).
Shearing force that occurs in a short column is expressed as the ratio of the sum of moments at the top and bottom of the short column to the height of short column (Eq.1).
M a and M ü are the moments of the top and bottom parts of the column where the wall is not placed, and l k is the length of the short column.As shown in Eq. ( 1), the shearing force in a short column is inversely proportional to the length of the short column, l k .Accordingly, when l k is short, the shearing force is high.The height of the short column is generally kept to 40-50 cm for lighting applications.Shearing forces due to earthquakes may cause great damage if adequate preventions are not taken in the short columns.The walls are generally not considered in the model during the project stage.In other words, the contribution of the rigidities of the infilled walls to structural behavior in the design stage is not taken into consideration.However, the rigidity of these elements is expected to affect structural behavior during an earthquake if the structure has both symmetric and asymmetric plans (Güney, D. and Boduroglu, 2006).
The following provisions and explanations related to short columns are given in the Turkish Earthquake Code.Short columns may be formed due to the bearing system or due to voids left between columns in infilled walls.If the formation of short columns is not prevented, the shearing force that will be experienced can be calculated as a function of lateral reinforcement as follows: The moments in Eq. ( 2) are calculated at the bottom and top of the short column as M a ∼ = 1.4M ra and M ü ∼ = 1.4M rü , and  l n is taken to be the length of the short column.However, the calculated shearing force will provide the conditions given in Eqs. ( 2) and (3).Throughout the short column, the minimum lateral reinforcement and settlement conditions described by the confinement zones of columns in TEC 2007 are applied.The lateral reinforcements are maintained for story height in columns that become short columns by remaining between infilled walls (Fig. 5).
Examples of damage that occurred in reinforced concrete buildings due to window bays left in infilled walls and column edges are shown in Figs. 6 and 7.

Damages depending on soft-story-weak-story irregularities
Some geometric and structural configurations addressed in the TEC 2007 in 2007 are described as irregularities; it was determined necessary to avoid the design and construction of irregular buildings due to the negative behaviors of these building during earthquakes.Irregularities are grouped into two categories: plan and verticality irregularities.There are three types of irregularities observed in verticality: strength irregularity between adjacent floors (weak floor -B1), rigid- ity irregularity between adjacent floors (soft story -B2) and discontinuity of vertical elements of load-bearing systems (Aktürk, 2004).
For one of two perpendicular earthquake directions in a reinforced concrete building, the rate of average relative displacement (( i ) average ) of any floor to that of the upper floor (( i+1 ) average ) is defined as the rigidity disorder coefficient (η ki ).When this coefficient is greater than 1.50, a B2 type irregularity is present in the structure (Isik, 2006).
In the calculation of average relative floor displacement (( i ) average ), the lateral earthquake load must be applied with a ± 5 % additional eccentricity.That means the distance (e) between the applied earthquake load and center of mass must be In this equation, L is the size of the structure perpendicular to the earthquake direction.Equation ( 4) should be calculated for both the x and y directions, which are perpendicular to each other and belong to the i th and (i + 1) th floors of the building.An operation should be carried out according to greater value.
According to TEC 2007, to determine whether there is a floor irregularity in a building, an η c criterion is applied.The criterion is described as follows, and an operation should be carried out according to the minimum value of η c : Here, for the earthquake direction under consideration, the following parameters are defined: A ef,i is the sum of the effective cross-sectional areas on a floor; A c is the sum of the cross-sectional areas of columns on a floor; A s is the sum of  the cross-sectional areas of partitions on a floor; A m is the sum of the cross-sectional areas of infilled walls on a floor.The above operations are repeated for the x and y directions of the building, and an operation is carried out according to Table 1 by considering the minimum η c .
When damage to buildings caused by earthquakes has been investigated, it was found that the buildings that had few or no masonry-infilled walls, typically on the ground floor, experienced greater damage at the ground floor than on the upper floors (Tezcan et al., 2007).Masonry-infilled walls are definitely not considered by engineers in the internal force calculations for reinforced concrete load-bearing systems.Because masonry-infilled walls are not allowed to minimize the inner cross-section demands of the main loadbearing system elements such as column-beam-partitions by taking a share of the horizontal earthquake load, the designs for load-bearing systems are less than adequate.However, masonry-infilled walls play a large role in decreasing the horizontal displacements of the floor on which they are installed.The resistance of a ground floor lacking infilled walls against horizontal displacement is less than the resistance of upper floors that have many infilled walls.For this reason, a floor that has a rigidity discontinuity in the horizontal direction is called a soft story.If the height of the ground floor is greater than the upper floors, it creates a soft-story irregularity.In order to provide space for the various commercial functions of shops, restaurants and banks, the ground floors that do not have infilled walls and/or have relatively high floor height are a focal point of earthquake damage in multi-story buildings (Tezcan et al., 2007).
Damage mainly occurs in a soft story when there are masonry-infilled walls on the upper floors but not on the ground floor.Because earthquake damage is caused by the lack of masonry-infilled walls, it is necessary to include masonry walls in the models used to calculate horizontal displacement.Otherwise earthquake damage might occur unexpectedly due to excess horizontal displacement in a weak story.Accordingly, the rigidity provided by infilled walls should absolutely be taken into consideration during the calculation of horizontal displacement and especially during the determination of the elastic first natural pulse period.
Plastic hinges in soft-story behavior occur when the lateral translation rigidity in the upper floors above the ground floor is greater and when the lateral rigidity of the ground floor is low.Horizontal earthquakes force cause great strain on the ground floor and increase the lateral displacement, causing plastic hinges to form.The structure cannot benefit from the extra rigidity provided by infilled walls, and an inconvenient situation develops that negatively affects the strength and translation of the ground floor.When non-load-bearing walls are not installed in the structure, the growing lateral translations will impair structure stability.When there is a discontinuity in the shear walls and these regions are necessary for construction, transition zones should be reinforced adequately.When the ground floor rigidity is lower than that of other floors, large increases are observed in the strength of structure when horizontal loads are applied.Lateral translations increase in the structure, and plastic hinges occur in the columns.Sudden rigidity changes in the ground floor and non-elastic behavior can cause severe damage at the top points of the ground floor columns, and such damage is detrimental according to earthquake-resistant structure principles.Moreover, large lateral translations that occur in ground floor columns will cause second-degree moments (Isik, 2006).
In Fig. 8, many examples are shown of soft-story and weak-story damages observed in the 2011 Van earthquake.
Many buildings were heavily damaged in the 17 August 1999 Kocaeli earthquake (M w : 7.4).Those buildings did not have masonry-infilled walls on their ground floors in order to leave space for commercial uses on Izmit and Golcuk Street and especially on AdapazarıInonu and Cark Street.The upper floors of the incomplete five-floor building shown in Fig. 9 were designed as residential floors, and the ground floor was designed for shops.This building, in which the ground floor was quite high and lacked masonry-infilled walls, was heavily damaged and collapsed due to weak-story syndrome.
According to the η c weak-story and η k soft-story criteria in the TEC 2007, weak-story and soft-story irregularities were not found in this building.Instead, the building collapsed due  to the weakness and softness of the ground floor.In this case, the TEC 2007 could not identify a weak or soft story in the building.Thus, the design engineer and the assignors were misled as to the threat to human life because the building was viewed through rose-colored glasses.

Damage caused by reinforcement detailing mistakes
Reinforcement detailing mistakes were frequently encountered in structures, especially those constructed before the establishment of the 1998 Earthquake Code.After the 2011 Van earthquake, typical reinforcement defects were observed in most of the heavily damaged buildings.Discontinuities in column lateral reinforcements in the joint region and inadequate length of the reinforcement interlocking sections were the most common detailing mistakes.
It was observed that hooks were attached to the ends of the longitudinal reinforcements in old reinforced concrete structures.In Fig. 10, a heavily damaged column element can be seen.The building, which was used as a primary school, contained many application and detailing defects.Beam lateral reinforcements were applied adjacent to the end region, and they were not integrated with concrete.Hooks were attached to the ends of the beam longitudinal reinforcements and continued at the end region, except for the lateral reinforcement.Column lateral reinforcements did not continue at the joint region of the column and beam.Moreover, the interlocking length of the column longitudinal reinforcement was inadequate.For these reasons, the reinforcements were pulled away from the joint region (Caycı, 2012).(Isik G., 2006).
According to the earthquake code, non-deformed reinforcement steel cannot be used except for stirrup and tie reinforcements that are combined with slab reinforcement.This decision was accepted in earthquake regulations in 1998.Non-deformed construction steel was used in many of the old buildings in Turkey's structural stock that were constructed before the 1998 earthquake code.The adherence of non-deformed steel with concrete is weak.Under reversiblerepeatable loads, adherence is required by reinforcements that are consistently exposed to tensile stress and compressive stress to transfer the load to the concrete.After losing adherence, reinforcement steel will be pulled away from the concrete, and the column or beam elements will be removed, especially from the column-beam joint points (Fig. 11).
Figure 10.A reinforced concrete structure that was heavily damaged at the joint region during the 19 May 2011 Simav earthquake (Caycı, 2012).
According to the TS (Turkish Standarts) 500-2000, the lap length (generally the development length) is determined by the quality of the steel and concrete used.The lap length for ST 420 steel and C20 concrete (or development length) should not be less than approximately 62 times the steel diameter.Accordingly, the lap length for ∅14 steel should be at least 90 cm, and for ∅16 steel it should be at least 100 cm.For ∅18 steel it should be at least 115 cm.Despite this, in applications the lap lengths are kept by disregarding the above rule and minimizing the amount of steel used.This weakens the joints of columns between floors.Thus, the columns break off during earthquakes when they are connected to sub-floors, and the building collapses.
According to the TEC 2007, special earthquake stirrups and special earthquake ties must be rearranged in beam and column confinement zones near beam-column joints.Column stirrups must be continued through the columns within the beam-column joints.Stirrup hooks that satisfy the earthquake code are shown in Fig. 12.
At both ends of the special earthquake stirrups, corrugated hooks with a 135 • angle must be present.The length of hooks must not be less than 10∅ and 100 mm in undeformed bars and not less than 6∅ and 80 mm in deformed bars, as measured from the last point of tangency.Making stirrup hooks with 90 • angles is straightforward.Bending stirrup hooks to an angle of 135 • increases the production time and labor costs.Because the importance of the correct angle has not been well understood, defective manufacturing has been very common.
A concrete core wall under axial pressure is forced to undergo lateral deformation due to the Poisson effect.The stirrups resist the lateral deformation of the concrete core wall by using longitudinal reinforcements as structural bearings.The prevention of lateral translation results in an increase in compressive strength in the concrete core wall.Moreover, the ductility of elements also increases significantly.
Another contribution of stirrups to element strength is that they shorten the bending length of the longitudinal reinforcement.Bending load will increase due to shortening of the bending length of the longitudinal reinforcements under axial pressure.
When stirrup hooks are made with a 90 • angle, the reinforcement hook only bonds with the cover concrete.The cover concrete first flakes off by cracking because it is not exposed to the confinement effect under repeatable loads.In this case, the stirrup hook will easily open outwards because the core concrete confinement of the stirrup is removed.The opening of the stirrup will result in loss of the confinement effect on the concrete core, which is forced to expand outwards.Moreover, because the buckling length of the longi- tudinal reinforcement is increased, buckling in the reinforcements can easily occur.Element ductility and bearing capacity will decrease.In Fig. 13, various examples are given for the opening of stirrups and the buckling of longitudinal reinforcements.
According to the TEC, special confinement zones must be placed at the bottom and top of each column.The length of each confinement zone must not be less than the maximum column cross section, one-sixth of the column clear headroom or 500 mm.In the confinement zones, lateral reinforcements with less than 8 diameter must not be used.In this zone, the stirrup range in the longitudinal direction must not be more than one-third of the smallest cross-sectional dimension or more than 100 mm, and it must not be less than 50 mm.
In columns with stirrups, the minimum total lateral reinforcement area in the confinement zone should be calculated based on the given conditions in Eqs. ( 8) and ( 9) when N d > 0.20 A c × f ck in this calculation, the core size of the column b k is taken into consideration separately for each direction, s is the longitudinal reinforcement interval, spiral reinforcement step, b k is the core size of column; A c is the gross crosssectional area of column or partition end zone; A ck is the column concrete core wall area within the measurement taken, out-to-out of confinement reinforcement; f ywk is the characteristic yield strength of lateral reinforcement; f ck is the characteristic cylinder compressive strength of concrete.
In the zone between the confinement zones at the bottom and top of the column (the middle zone of column), a lateral reinforcement with less than ∅8 diameter must not be used.A stirrup, tie or spiral interval throughout the column must not be more than half of the smallest cross-sectional dimension or more than 200 mm.The horizontal distance between stirrup levers and/or ties, a, must not be more than 25 times the diameter of the stirrup.In the TEC 2007, although confinement zones are defined at the end points of columns and beams and it was determined that the stirrup frequency in the confinement zones should not be less than 10 cm, it was observed that such confinement zones were not formed in reinforced concrete elements damaged in the Van earthquake (Fig. 13).It was observed that the stirrup frequency at the end zones of the columns and beams was approximately 20-30 cm.Plastic hinges are formed at the column ends due to strong-beamweak-column formation under reversible loads.Because of the infrequent stirrups under excess spinning caused by the ductility request and because hooks with 90 • adequate ductility were not obtained, correspondingly heavy damage was observed in many of the buildings.Inadequate concrete quality also contributed to these failures.In Fig. 14, examples of column damage and stirrup frequency are given.
The most constraining zones within the framework under earthquake loads are column-beam joint regions.These zones must withstand both the reversible moment and the axial load resulting from the column and reversible moments that the connected beams are exposed to.The increased frequency of stirrups used in the confinement zones of columns in these zones must be continued exactly.Moreover, columns and beams connected in this zone should provide enough overlapping length to provide longitudinal reinforcements.The decisions related to overlapping length for longitudinal reinforcements at joint region in the TEC 2007 are given in Fig. 15.
The stirrups in the confinement zones of columns should be continued with the same frequency in the column-beam joint zone.These zones are the most important ones in terms of both the bending moment and the shearing forces under reversible-repeatable loads.The stirrups increase the spin capacity of the element joint by increasing ductility and contribute to shear strength.However, the continuity of column stirrups in joint regions are difficult to maintain during manufacturing, and it is a detail that has not been included until recently.Most people do not understand the importance of it.Examples of damage in column-beam joint regions are given in Fig. 16.

Damage caused by weak-column-strong-beam issues
All modern earthquake codes specify that columns should be stronger than beams.Thus, the plastic hinges under horizontal earthquake loads will form first on beams and then on the columns at the same node.Before system stability is lost, the number of plastic hinges formed on beams is greater than the total number of hinges formed by the same hinges on the column.Thus, more hinges form in the system, and the ductility of the system increases due to greater translation of the system and a greater amount of energy consumed.Moreover, the axial load-bearing level of the beams is lower than that of columns, so the ductility of the beams under bending moments will be more than that of columns.Thus, the stabil- Figure 19.Full collapse of a building as a result of plastic hinges in the bottom and top zones of columns in a reinforced concrete structure in Van-Ercis (Onen et al., 2011).
ity loss of the columns will cause the collapse of the structure.However, even when hinges form in all of the beams, the building may not collapse.Repair or reinforcement after an earthquake by adding hinges to columns is difficult and may be economically impossible, but repair and reinforcement of a damaged building by adding hinges to beams is easier.For these reasons, the TEC 2007 states that "in a loadbearing system consisting of only frameworks or a combination of partitions and frameworks, the total bearing capacity moments of the columns integrating at each column-beam node point (Fig. 17) must be at least 20 % more than the total of the bearing capacity moments in cross sections of the column surfaces of beams integrated at that node point".
M ra is the bearing capacity moment calculated by taking f cd (concrete design compressive strength) and f yd (steel design tensile strength) of the column under the joint as a basis.M rü is the bearing capacity moment calculated by taking f cd and f yd of the column above the node point as a basis.M ri and M rj are the bearing capacity moments calculated by taking f cd and f yd at the beam ends as a basis.This equation will be applied separately in each earthquake direction.In the calculation of column-bearing capacity moments, N d axial forces that result in the smallest moments for the direction of the earthquake will be taken into consideration.
The reason for the condition determined in Eq. ( 10) is that beams should be more ductile than columns because ductility decreases as axial load increases.By taking this reality into consideration in the code, it is possible to ensure the formation of plastic hinging on more ductile beams, which is unavoidable in big earthquakes (increase in curvature with tension reinforcement yielding under constant moment).This condition is determined by Eq. ( 10) as "columns should be stronger than beams" (Ersoy and Oczebe, 2004).
Although it was stated in the earthquake code that columns should be stronger than beams and that hinging occurred in columns investigated after earthquakes, it was observed that the damage to beams was limited and that in some cases there was no damage.Several examples of this type of damage, which is called strong-beam-weak-column damage, are given in Fig. 18.As the damage becomes progressively worse, hinges occur in the columns, decreasing the horizontal stability and causing the building to collapse in a way that floors collapse one after another, which is called a "pancake" collapse.Pancake collapses are the most common type of collapse (Fig. 19).
One of the most commonly encountered problems in damage investigations after earthquakes is that the buildings do not have adequate lateral rigidity or an adequate layout of shear walls.Ductile behavior is desired in a structure under earthquake loads, and all structure systems should have adequate rigidity.More horizontal translation will cause damage to elements such as non-load-bearing partition walls, and this will increase the economic losses.Moreover, second-degree forces will cause more movement of the elements.For these reasons, the TEC 2007 imposes restrictions on the horizontal displacements that occur in floors under horizontal loads.
For each earthquake direction, the maximum value (δ i ) of calculated effective relative floor translation in columns or partitions on the i th floor of a building must satisfy the following condition: If this condition is not satisfied on any of the floors of a building, the earthquake calculation needs to be repeated by increasing the rigidity of the load-bearing system.However, even if the condition is satisfied, the use of nonstructural brittle elements (frontal elements, etc.) under effective relative floor translation needs to be confirmed by the calculations.Buildings must have adequate shear walls in both directions in order to have adequate lateral rigidity.However, the construction of shear walls is not sufficient to rehabilitate old buildings.The building shown in Fig. 20 was constructed near Van Lake.
Figure 21 shows examples of earthquake damage and collapse observed in reinforced concrete buildings with insufficient lateral rigidity.
Figure 22 shows examples of damage in non-bearing brick-infilled walls as a result of excess translation of buildings.
Walls with upper sides that are clear and supported at the bottom, such as vertical corbels or garden walls, frequently collapse under horizontal loads because the gable walls are not combined with overhead beams in garrets.The collapse of these walls does not pose a problem in terms of the veering of the building.However, the collapse of these walls on the street or in other buildings might put people in danger.Figure 23 shows examples of gable wall damage.
In the buildings located near to the main streets, damage occurred due to the placement of vertical load-bearing system elements.All columns in the building are positioned vertically relative to the street, and there are no columns parallel to the street.It is difficult to know the direction of the earthquake that affected the building, but the component of the earthquake that is parallel to the street caused heavy damage in the building.

Rigidity
A structure must have adequate rigidity to minimize the second-degree moments as much as possible and to reduce the nonstructural damage in earthquakes that correspond to the serviceability limit state.The most important criteria for structural rigidity under the effect of horizontal loads are the rigidity of the element and the translation of the floor relative to a lower floor in the structure.For rigidity, the positions of the vertical bearings and their dimensions in both directions are very important, more so than the geometry of the structure.In a reinforced concrete structure, because the columns and partitions behave more rigidly as vertical load-bearing elements than infilled walls, it is sufficient to take these elements into consideration in the calculation.In the calculation of rigidity for vertical loadbearing elements, the properties of the materials used in the elements (elasticity modulus), the cross-sectional dimensions of the element and the supporting types of element ends must be considered.
Relative floor translation is related to displacements that might occur in an earthquake, and the structural rigidity in the earthquake code defines the displacement difference between two floors as i = d İ − (d i − 1).For one of the two perpendicular earthquake directions, if the ratio of the maximum relative floor translation in any floor to the relative translation in the same direction in that floor (which is defined as the torsion irregularity coefficient) is greater than 1.2, then it is described as a torsion irregularity.To prevent torsion irregularities, attention should be paid to the displacements of partitions and rigid load-bearing elements in order to increase the torsional rigidity of the system (Aktan and Kirac, 2010).

Displacement of partitions in the plan
In reinforced concrete structures, a torsional moment occurs due to the geometry of the system in constructional elements or non-symmetrical loadings.The torsional rigidity of the system changes according to the cross section of the partition and its displacement in the plan.
The partitions should have buckling stability and symmetric rigidity and should be secure against toppling in the basement.When placing partitions in the plan, it is important to provide the distribution of expected plastic strains accurately in the building plan.Otherwise some partitions will be forced more, and some of them will be forced beyond their capacities.In order to provide adequate rigidity in a highly partitioned construction, at least three partitions should be used in which the system lines do not pass on a point.In order to eliminate torsion formation in a building experiencing an earthquake, a consistency between mass and rigidity centers is required.Earthquake forces affect the center of mass in a floor.If the center of mass does not conflict with the rigidity center, the construction will turn around the center of rigidity.When the horizontal force affecting the center of mass is moved to the center of rigidity, as is the F y force, a torsional moment with a value of M z = F ye will affect the center of rigidity.In fact, if the earthquake load affects the center of rigidity directly, the torsional moment will not occur because the construction will produce equal translation in the direction of the earthquake force.Because the torsional effect on the floor is the product of the shearing force affecting the vertical elements and the moment arm, the torsional effect in the partition or framework with the largest moment arm will be higher.To decrease the torsional effect on partitions, it is necessary to arrange the partition systems in an ideal way.This is done as follows.
1.In order to provide maximum torsional rigidity in the construction, shear walls should be distributed around the construction.In Fig. 24, suitable partition placements with high torsional rigidity are shown.
2. Shear walls should be arranged in the floor plan so that most of the floor loads are transferred to the basis as axial force.In this way, the reinforcement required for the bending moment in the partition is reduced.
3. In multi-story buildings, putting the earthquake resistance into few partitions exposes the basis system to a big earthquake effect at a few points.This situation should be avoided because it requires an expensive heavy basis system.
.6 Damage that depends on the hammering effect

Earthquake joint
Many buildings by law allow adjoined constructions (attached).These attached constructions are not well protected against earthquakes.They transfer earthquake force to each  other, and they collide as a result of oscillation.The outermost such construction, which typically occurs at a street corner, is generally damaged very heavily (Turk, 2011).
A new building that is built next to an old one is separated with a joint.There is an attempt to prevent collisions of the two buildings due to different oscillations in the earthquake, which would cause the buildings to damage each other (the doubling effect).When a new construction site has bossage or big spaces in the plan and/or in the vertical direction, it is deblocked into symmetric and rectangular spaces as much as possible.Blocks are constructed with a joint between them.Theoretically, the joint width between blocks should be greater than the sum of the maximum horizontal displacements of both constructions: If one of the constructions is old, its displacement is generally not known.In this case, it is recommended to take the joint width as a minimum: Here, H is the height of the structure.As an example, if H = 15 m, then the joint width in the construction will be at least The foundations of separate building blocks also should have separation joint (Turk, 2011).
Figure 25 shows examples of the types of damage that occur when there is not an adequate earthquake joint between attached buildings.

.7 Damage observed in non-load-bearing construction elements
The construction elements that do not have load-bearing properties in buildings but cause maximum damage during an earthquake are the partition walls.Partition walls are built to separate the usage areas physically within a building and to protect internal volumes in the edge axis from the outside, and they are generally made of hollow bricks in Turkey.However, the outer walls that are called sandwiched walls due to climatic conditions in East Anatolia Region are made of two-row hollow bricks, generally with polystyrene foam and glass wool between the bricks to provide heat insulation.
In two-row brick walls that are not connected to each other mechanically throughout the wall gaps, diagonal cracks and large out-of-plane motions have been observed.For this reason, there have been many cases of buildings with no damage in load-bearing components but with severe loss of life and property due to wall damage (Fig. 26) (Celebi et al., 2011).
.8 Damage depending on the suitability of materials used in load-bearing elements In the concrete that is generally used for construction, segregations and reinforcement placement defects are observed.Such defects were observed in all investigated buildings (Fig. 27).20) and a concrete cross section displayed in the wreckage (Kizilkanat et al., 2011).

Discussions and conclusions
The most important causes of damage during earthquakes have to do with a building not being constructed according to modern codes and standards.
The only way to construct earthquake-resistant buildings is to be aware of the requirements and to avoid the mistakes of the past.It is necessary to have good communication and collaboration between architects, civil engineers, geologists, urban and regional planners and related professional organizations for the construction of earthquake-resistant buildings.
It is impossible to predict earthquakes, but it is possible to construct earthquake-resistant buildings.Visual values such as aesthetics and artistic value should be secondary to safety in a construction project.
The construction of earthquake-resistant buildings is the joint responsibility of engineers and architects, and it necessitates the cooperation and systematic study of related professions.
The most important thing during an earthquake is the strength of the building.For this reason, no matter how perfect the calculations are, when the manufacturing and construction processes are not performed carefully the building will not perform as expected in the event of an earthquake.Thus, it is necessary to instill the control mechanisms that are currently lacking in our country as soon as possible.
Because earthquakes affect buildings as horizontal loads, an adequate number of shear walls should be placed to increase the lateral rigidity and decrease translation during the construction of a load-bearing system.Architects and engineers should investigate and understand the factors that cause damage in earthquakes.
In the education of architects and engineers, earthquake and earthquake-resistant construction design should be considered important, and previous graduates should have their education updated via seminars or courses.
In Turkey, what is said after every earthquake is the same as what was said after previous earthquakes.The defects and mistakes that cause damage do not change.If we do not want to experience the same negative consequences, we should identify the problems correctly and take the necessary precautions.
Three causes of earthquake damage can be identified: -design mistakes (soft story, inadequate lateral rigidity, short column, strong-beam-weak-column joints, irregularities in vertical and horizontal directions, etc.) -construction stage mistakes (poor workmanship, low strength of materials, inadequate transverse reinforcement-stirrup usage, defective and inadequate interlocking length, lack of control) -mistakes in usage (cutting column, soft-story formation by complete or partial removal of walls, etc.).
In the strongest earthquakes, especially those that cause permanent damage to buildings, it is necessary to obey the criteria given in codes for the controlled distribution of damage in construction in order to avoid a collapse.The following suggestions should be taken into consideration in construction to avoid damage and collapse: a.The construction system selected should be as simple as possible and should be a system that can be easily understood by everybody involved in the project.
b. Various load-bearing components of the building may have the property of sharing horizontal loads.
c. Plastic deformations should be received by all components without loss of stability under vertical and horizontal loads.A capacity design should be used, and the maximum horizontal translation should be enhanced within the limits given in code.
When a designer pays attention to these details, the required conditions can generally be provided for the elastic and ductile behavior of a building.
Edited by: D. Keefer Reviewed by: three anonymous referees
Turkey during the last 25 years and the types of damage observed in reinforced concrete buildings after the 2011 Van earthquake

Figure 4 .
Figure 4. Short column behavior and formations in structures due to various causes.

Figure 6 .
Figure 6.Examples of damage caused by the formation of short columns (METU/EERC, 2011).

Figure 8 .
Figure 8. Examples of damage related to soft-story and weak-story irregularities (source: Ahmet Topçu).

Figure 14 .
Figure14.Damage caused by excess stirrup intervals in the confinement zones of columns(Onen et al., 2011).

Figure 20 .
Figure 20.Buildings that do not have adequate shear walls.

Figure 21 .
Figure 21.Damage observed in buildings with insufficient lateral rigidity.

Figure 25 .
Figure25.Reinforced concrete multi-story building that collapsed by colliding with adjacent buildings due to a hammering effect(Inel et al., 2011).

Figure 26 .
Figure 26.Damage in the outer walls of a reinforced concrete building in Van city due to separations in the walls (Cagatay I.H., 2007).

Figure 27 .
Figure 27.Concrete and aggregate samples in a single-story reinforced concrete building that collapsed in the Van organized industrial site (20) and a concrete cross section displayed in the wreckage(Kizilkanat et al., 2011).

Table 1 .
Operations according to the weak-story criterion η c .