The Working Group 4 "Prefabricated Building Structures in Seismic Regions" was formed in 1986 at the 8th European Conference on Earthquake Engineering carried out in Lisbon, Portugal.

The present report is the 3rd one about the activity of the Working Group. This report summarises the work during the last several years of the specialists from four countries: Bulgaria, Former Yugoslavian Republic Macedonia, Russia and Turkey.

The investigations of 19 Bulgarian specialists are reflected in the report. Some theoretical and experimental investigations performed in Skopje were presented by Working Group Participant Dr.Z.Bozinovski. Prof. J.Eisenberg from Russia sent a material about Prefabricated Building in his country. The Turkish participant Prof.Ersoy prepared a document summarising the application and research on seismic resistant precast concrete structures in Turkey.

As the convenor of this WG4 I would like to express my gratitude to all colleagues, who took part either directly or indirectly in the WG activity. Special thanks to the Bulgarian colleagues, research associates M.Kouteva and A.Kaneva, WG participants, who helped me in writing the final report.

In spite of the fact that due to financial reasons WG meeting could not be held to discuss the contents I hope that this report will be useful for specialists related to the Earthquake Engineering.

Prof.Dr.D.Nenov

W.G. Convenor

At its meeting in Skopje, September 1983 the EAEE Executive Committee decided to organise ten working groups entrusting them to prepare reports on some important problems of mutual interest for the 8th European Conference on Earthquake Engineering in Portugal, September 1986.

Bulgaria (Prof.D.Nenov) was appointed to be responsible for WG 3 "Prefabricated Large-Panel Buildings in Seismic Regions". The WG members were 12 specialists: 8 from Bulgaria, 2 from Yugoslavia, 1 from Romania, 1 from CSSR. The Bulgarian National Committee had ensured the necessary financial support and a WG meeting was held in Sofia during 27- 29 November 1985.

The final WG report was published in the "Proceedings of the 8th European Conference on Earthquake Engineering", Volume 7, pp.73-124, Lisbon, 1986. The titles of the sections included in the report were as follows:

1. Review of the international experience on prefabricated large-panel building construction in seismic regions
2. Connections of the large-panel buildings
3. Analysis of standard documents for design and construction of large-panel buildings in seismic regions
4. Analysis of large panel buildings behaviour during earthquakes
5. Theoretical and experimental investigations related to the seismic safety of the large-panel buildings
6. Some new trends in the development of large-panel buildings in seismic regions
7. Concluding notes and some recommendations.

The Working Group 4 "Prefabricated Building Structures in Seismic Regions" is one of the eight working groups formed in 1986 at the 8th European Conference on Earthquake Engineering carried out in Lisbon, Portugal, with convenor Prof. D. Nenov. Specialists from 5 countries took part in the work (from Bulgaria, Italy, Romania, former USSR and Yugoslavia). The Bulgarian National Committee for Earthquake Engineering had ensured the necessary financial support and a WG meeting was held in Sofia from 14 to 17 November 1988.

The final report was published in the "Proceedings of the Ninth European Conference on Earthquake Engineering", Volume A, pp.145-204, Moscow, 1990. The titles of the sections included in the report were as follows:

1. Generally about the prefabricated buildings
2. Design experience and application in some European countries
3. Requirements in the design and construction of the prefabricated building structures in seismic regions
4. Joints and connections
5. Analysis of the prefabricated building structures’ behaviour during earthquakes
6. Some new trends in the development of the prefabricated building structures in seismic regions.

More details concerning the prefabricated building construction in Bulgaria, Italy, Romania, USSR and Yugoslavia are given in the six annexes to the report.

During the interval 1990-1994 the Working Group 4 "Prefabricated Building Structures in Seismic Regions" continued to work. The participants were from three countries: Bulgaria, Former Yugoslavian republic of Macedonia and Turkey. Because of the lack of financial support a meeting of the participants in the group could not be organised.

The final report was published in the "Proceedings of the Tenth European Conference on Earthquake Engineering", pp.3079-3097, Vienna, 1994. In this report it was emphasised on some specific problems which are the object of the recent investigations of researchers and specialists. The titles of the sections included in the report were as follows:

For the period 1994-1998 the WG 4 "Prefabricated Building Structures in Seismic Regions" continues to work on some special engineering problems. The results of these investigations are given below.

Whatever catalogue used for the prefabricated buildings, the problem of the connection between the elements is of crucial importance. By means of the connections the different elements form the structure of the building which must bear the effects for which it is designed. Hence the great importance of the connections, especially for constructions built in seismic regions. On the other hand the joining of the different elements is not performed in the factories but on the building site where conditions can vary due to atmospheric and other reasons. That is why it is reasonable to investigate the connections particularly in buildings, which were longer used. For that purpose connections in buildings in exploitation had to be uncovered and investigated. The investigation presented by Nenov (1997), included altogether 12 prefabricated buildings constructed in Sofia, Plovdiv, Russe and Burgas. The total number of the investigated connections is 29. Some of the most important findings are:

• In the majority of unrecovered connections there are not serious deviations in the geometric dimensions and the metal parts in them in comparison with the requirements of the designs;
• All weldings are performed in correspondence to the design or with deviations which do not endanger the bearing capacity of the connections;
• In most cases the infilling concrete is with great strength and well envelopes the joining steel bars and no gaps are noticed between this concrete and the panel concrete. This indicates that the good infilling of the connections with concrete is completely possible. Unfortunately during the infilling of some connections serious deviations from the design projects are made: concrete of poor quality has been used and in consequence of it gaps between the infilling concrete and the panels concrete have been formed; in the other cases instead of using a cement mortar for the joining steel parts, they are covered with lime mortar. Obviously such failings diminish the durability of the connection and they must not be tolerated;
• Different degrees of carbonisation of the infilling concrete are found depending both on the exploitation periods and the microclimate in the different buildings and premises. In the building with a term of exploitation about 10 years the carbonisation is still small and affects predominantly the surface layer of the concrete. In other buildings of about 15 years of exploitation this carbonisation is more advanced. In one case a complete carbonisation was noted. This means the concrete protecting properties against the corrosion of the metal parts have been significantly decreased.
• Concerning the corrosion of the metal parts the following can be pointed out:

All discovered steel links in the connections when the latter are realised in correspondence with the design requirements are reliably protected against corrosion. This also refers to the cases when they have not an anticorrosion covering. In these cases the metal parts are without or with insignificant rust. The velocity of the corrosion is of order of 0.01 to 0.001mm per year, without affecting the metal structure.

In spite of this fact the right thing to do is not to neglect the anticorrosion covering over the metal links. On one hand some omissions in the filling of the connections with concrete cannot totally be avoided, and on the other hand it must be born in mind that in the course of time the protecting properties of the concrete decrease. The investigations established that when there is anticorrosion covering on the steel parts there is no corrosion.

In the connections in which between the infilling concrete and the panels concrete gaps have been formed and the metal parts are not covered with concrete, the corrosion on them is well expressed. If such connections are situated in wet premises there is danger of a fast corrosion. In the course of time steel parts can significantly decrease their cross section and the bearing capacity of the connection will impermissibly diminish before the exploitation term of the building has expired.

The above findings lead to generalised conclusion that the designers, investors and dwellers of the prefabricated buildings should not worry about the reliability of this type of buildings provided the connections in these buildings are made in correspondence with the design.

On the other hand the defects made during the performance of the connections is a serious and alarming signal, which requires undertaking the corresponding measures.

One of these measures is: the concrete and steel corrosion in the connections, and specially the concrete carbonisation, to be constantly (periodically) inspected and studied in order the latter to be stopped or diminished in due time.

As a general rule it is to be said that the concrete corrosion is possible when an aggressive environment exists. The environment is aggressive when there is acids in it, either in the form of water solution or as acid gas. The environment can be aggressive too when there are some salts or even alkalis in it. Except the chemical composition of the environment many other factors influence on the concrete corrosion speed: cement quality and quantity, concrete quality, stress state, cracks appearance, temperature, etc. Nenov et al. (1998).

Figure 1. Arrangement of the salts in the concrete, formed by the influence of different gas; I-zone of reaction with SO₂, HCl, Cl₂; II-carbonised layer; III-not carbonised layer.

The combined action of the different acid gas leads to the creation of different salts which are situated as it is shown in Figure 1, Moskvin et al. (1980).

The analysis of the combined acid gas action indicates that the dominating influence on the concrete neutralisation falls to the carbon dioxide (CO₂). The other gas accelerates or decreases to a certain degree this process. That is why special attention is paid to the carbonisation.

In the case when the thickness of the carbonated concrete for the exploitation term is smaller than the concrete cover, the structure could work without special covering against the corrosion. Such conclusion can be done if a method for determination of the carbonisation speed is established.

Specialists from different countries work on the problem to prognosticate the carbonisation speed. As a result of these efforts the following formula is established (Moskvin et al. 1980):

where: x is carbonisation depth, cm; D’ is carbon dioxide (CO₂) effective diffusion coefficient in the concrete, cm²/day; C_o is CO₂ concentration in the air, relative value; m_o is concrete reactive capacity, relative value, that is to say, the CO₂ volume to one concrete volume; t is time in days.

Formulas similar to Equation 1 are established by other authors too. Equation 1 can be written by the following expression:

Equation 2 represents a family of straight lines in a coordinate system x - t^1/2 shown in Figure 2.

When the CO₂ concentration and the air humidity are constant each particular concrete will have its coefficient A_i:

that is to say each particular concrete has its straight line.

However A could be a constant for one family of lines with D’ / m_o = const but with different values of D’ and m_o. So one line corresponds to a family of concretes with D ‘/ m_o = const.

The boundaries D’ in this correlation are determined by

Giving extreme values of the cement quantity (C), 200 and 400 kg / m³, and the carbon dioxide (CO₂) content in the carbonated hardened cement paste (K), 0.2 and 0.4, the result for m_o and respectively D ‘ can differ, for a concrete family with A = const, not more than 4 times and carbonated depth - 2 times, respectively.

In Figure 2 each of the three sectors includes concretes with different diffusion permeability (D’ x 10^-4 = 0.2; 1 and 10).

Figure 2. Chart for determination of carbonisation depth of the concrete in air environment humidity and with different effective diffusion coefficients D’ of the carbon dioxide (CO₂).

The straight lines in Figure 2 represent the correlation x - t^1/2. Better representation of the x - t dependence is shown in Figure 3.

Figure 3.

Other investigations (Martin et al. 1975) prove that Equation 1 can be used only for premises with heating. For the constructions in the open air the following formula is proposed:

The greatest difficulty of the carbonatization prognostic using Equation 1 is the estimation of D’. That is why the most exact carbonatization can be established on the base of direct diffusion permeability estimation of particular concrete in its exploitation conditions. In this case two variants are possible. The first one is based on using the carbonated depth established for the several years of exploitation and the second one, based on the accelerating carbonatization of specimens in an increased CO₂ concentration.

The carbonatization depth x₂ after t₂ years can be established by Equation 6:

where x₁ - the depth of carbonatization for t₁ years exploitation; C₁ and C₂ - CO₂ concentration for the test samples and in the natural conditions, respectively (for the first variant C₁ = C₂).

Using the above methods the carbonisation depth in existing buildings can be established. This has to be done especially for the connections in prefabricated buildings in seismic region, because the durability of these connections is a determinant factor for the reliability of the construction.

That is why the observation about the carbonisation has to be a constant care.

When the result of this investigation shows that the carbonatization depth for the term of exploitation is smaller or equal to the thickness of the concrete covering on the steel parts, special measures against the concrete and steel corrosion are not needed.

But what to do when the protecting property of the concrete could be exhausted before the end of the exploitation term?

One radical solution in this case is the concrete in the connections to be removed and replaced by a new concrete. But this operation is not always suitable in a building of exploitation. That is why it has to be used when other measures are not possible.

The other solution is different protecting coverings to be used in order to stop or to diminish the corrosion process. The function of one covering is to not allow or to limit the possibilities of contact between the concrete and the aggressive environment, so that the connections to be reliable for the exploitation term. Specht (1991) in Figure 4 gives one example of this protection method.

The different protecting coverings could be waterproof materials, varnish materials, sticking materials, facing materials and so on.

The preparation of the concrete surface before the placing of the protecting materials is very important. The estimation of this problem concerns the humidity of the surface concrete layer, the surface roughness and cleanness. For example the humidity of the surface concrete layer with 20 mm thickness has not to be more than 5 - 6 %.

Figure 4. Parabolic and straight lines functions for concrete carbonatization without covering and with covering (with and without chalking) according to M.Specht.

This study is carried out by a team under the supervision of Prof.M.Dimitrov (1989).

In the Bulgarian code for design of buildings and structures in seismic regions (1987) it is prescribed the seismic forces to be determined depending on the structure ductility, i.e. on the structure capability to absorb energy allowing nonelastic deformations to be developed. According the code the "response coefficient" is assumed as ductility measure. It represents the possibility the construction to absorb and dissipate some of the energy transmitted by the soil during earthquake excitations. To determine this coefficient for the built up in Bulgaria large panel buildings some experimental tests on the ductility of the joints of large panel residential buildings are carried out. Considering this type of structures it is assumed that the horizontal forces are borne by continuos vertical diaphragms consisted of panels where joints are supposed to have enough capacity to bear the force and assure the continuity.

Full-scale study of vertical diaphragms built up by wall elements is performed. Two types of elements are tested - with and without opening in the vertical diaphragm. Both specimens consist of two parts - wall element and basement. The specimens are loaded by horizontal and vertical forces, modelling the real panel loading.

The obtained results coming from the experiments include: hysteretic curves at all loading steps, stresses in characteristic points of the concrete structure, deformations at single points, data about the loading when cracks are being developed and failure loading.

Details can be seen in Dimitrov et al. (1989). Here it will be emphasised on the following conclusions:

• The ductility of the investigated large panel residential system elements is rather high, which allows significant nonlinear deformations to be developed after first cracks occurrence.
• The values of the response coefficient, determined under the assumption for equal deformations are obtained to be the same - 0.25. The upper values under the assumption of same absorbed energy in the case of elastic and nonelastic element behaviour are: 0.234 for panel without opening, 0.226 for panel with opening. Considering the element behaviour under different loading ranges and the obtained corresponding hysteretic curves, more reasonable is to assume those values based on the assumption of equal dissipated energy.
• Dissipation energy coefficient is different for the different stages of the work of each element. It increases fast after first cracks occurrence. The coefficient value before the first crack can be assumed to be 0.27-0.30, but after the first cracks and for loading near to the failure one, the coefficient values are in the range of 0.55-0.60.

This study is carried out by a team under the supervision of Prof.M.Dimitrov (1995).

The full-scale testing of two identical residential five storeys buildings in Sofia (Ovcha kupel 2) constructed by large panel system has been carried out by the Building Research Institute. There is just one difference between the two tested buildings. The underground floor of the one is cast in situ reinforced concrete structure and the other one is constructed by precast surrounding reinforced concrete walls and internal cast in situ reinforced concrete walls. Both residential have five storeys and are based on band foundations. The purpose of this full scale tests is to determine the dynamic characteristics of both residential and to study their behaviour under quasidynamic characteristics simulated by horizontal forces, corresponding to the equivalent seismic excitations for regions with seismic coefficient 0.27 (IX MSK Intensity). The full-scale test has been performed in the following three stages:

1. Dynamic tests in order to determine the dynamic characteristics of each residential building before the quasidynamic loading;
2. Quasi-dynamic tests for determining the loading capacity and the deformability of both structures;
3. Dynamic tests in order to determine the dynamic characteristics after the quasi-dynamic tests.

The dynamic studies in the first and the third stages are carried out using identical loading schemes and same methodology. The most important characteristics of this methodology are as follows:

• The building vibrations are excited by servo controlled electro-hydraulic system MTS, USA. It is equipped by special dynamic cars with mobile inertial masses, fixed to the moving piston bar of the activator force cylinder;
• To register the vibrations excited in different points, eight channel measurement system is created. It is based on individual vibrometers SMU - HBM, Germany. Vibration analogue registration is done using light-ray 8LS-1 oscillograph (Honuwell, USA).

The dynamic loading applied to the fifth level (IVth floor) is realised by two exciting cars located as it is shown in Figure 5. Car No1 is oriented along the longitudinal structure direction and the second car is transversally oriented.

Figure 5. Situation of the dynamic loading cars.

These tests are based on two types of loading:

1. short term rectangular impulse with low frequency - up to 1Hz;
2. sinusoidal vibrations with different frequency and amplitude.

The following loading regimes are applied:

3. longitudinal vibrations, which excite mainly translational free vibrations modes. It is realised using only car No1;
4. translation and rotation free vibrations modes realised using only car No 2. The rotational vibrations are very low.
5. translation vibrations in both directions as well as rotation ones, due to the simultaneously loading realised using both of the cars.

Vibration registration is done using induction strain gauges, located at 1^st, 2^nd and 5^th floor levels.

The most important items related to the quasi-dynamic test are given below:

1. The quasi-dynamic test is performed using six powerful hydraulic cricks, placed at three different floor levels (above 1st, 2nd and 5th floors). Loading is realised in four cycles: 140kN per crick in the first cycle, 240kN in the second cycle, 420kN in the third cycle and 560kN per crick in the fourth cycle. Each cycle consists of sign varying loading according to four steps scheme. First step is loading from 0 to +N, second step is unloading from +N to 0, third step is loading from 0 to -N and the fourth step is unloading from -N to 0, where N is the applied loading in a single crick. During the loading and unloading processes equal forces are applied to all cricks.
2. Strains and displacements measurement during loading in each step of loading from 0 to N and backwards are done at each 35kN loading interval for the first and second cycle, and at each 70kN for the third and fourth cycle. Time duration of one cycle is about 22 to 25 min.
3. The absolute and relative displacements and rotations are determined using inductive strain gauges located at first, third and fifth floor level. All of the strain gauges are connected with registration computer system.

The following more important conclusions and observations based on these three stages experimental study and the analysis, which was carried out can be formulated:

• In both cases, before and after the quasidynamic loading, the free vibration frequencies of the section with precast ground floor are generally lower with 1-17% in comparison with the cast in situ ground floor. When the loading is applied in both directions, longitudinal and transversal, the frequency f_i varies from 5.20 to 27.5Hz for the section with precast ground floor and from 5.25 to 33.0Hz for the other one.
• The maximum horizontal displacements during the different cycles of quasidynamic loading at a given level of both sections are either equal or differ negligibly (up to 10%), higher values are related to the precast ground floor section. The horizontal displacements of the different floors increase in upwards direction in given loading cycle for both sections. The measured rotations of the sections are equal of the same levels during the same loading cycles or differ with less than 10%.
• The logarithmic damping decrement before and after the quasidynamic loading varies in large interval. In general it is higher for the precast ground floor section. The damping after quasidynamic loading is applied is higher than before loading. The difference is rather small and does not affect directly the structure behaviour.
• The energy absorption coefficient is lower with 1 to 9 % in the case of cast in situ ground floor. This difference is rather small and does not affect directly the behaviour of the pre-cast ground floor section. If such influence can be considered, it can be assessed as positive one.
• Both sections manifest similar failures after the quasidynamic loading is applied.

The data obtained by this experimental study, as well as the analysis, that is carried out, all conclusions and observations lead to the general conclusion for equal seismic safety of both sections, when they are exposed to seismic excitations corresponding to region with MSK intensity IX (seismic coefficient K_s = 0.27). The results show that in the case of large panel residential buildings with height up to five storeys, constructed in seismic regions with MSK intensity IX, it is possible that ground floors to be with precast outer walls and cast in situ inner walls. Such a ground floor can not affect the seismic safety of the structure, when it is founded under normal geological conditions.

Very interesting material on "Prefabricated Large Panel Reinforced Concrete Systems" was presented by Bozinovki (1997), IZIIS, Skopje. The most important sections of this summarising material concern some experimental investigations, proportioning of elements, nonlinear dynamic response of prefabricated systems exposed to dynamic seismic effects, design and analysis procedure as well as some conclusions and recommendations. As illustration some data about experimental investigations and nonlinear dynamic response are given below.

The experimental investigations refer to a structure of ground floor + twelve storeys with bearing walls thickness b = 16 cm. The main objectives of the experimental investigations are definition of the main strength and deformability characteristics of the vertical wall panels and proving of their ductile mechanism of behaviour.

Figure 6. Distribution of the equipment and position of the model during testing.

Details about the elements construction, characteristics of the built-in materials, models casting in situ, experiment performance equipment, instrumentation of the models, loading program, performance of experiment and results as well as their analysis can be seen in the material of Bozinovski (1997). Based on the analysis of the results from the experimental investigations, the more important conclusions are drawn:

• The panels are characterised by a ductile behaviour in all phases, i.e. from the first cracks occurrence in the panels tensile part next to the contact with the horizontal joint where the bending moment is the highest one, trough achieving the ultimate elastic limit of the reinforcement at the ends, crushing of concrete at the corners on a relatively small area, up to failure of the reinforcement at the ends;
• Damages of the model are concentrated around the initial crack, at the connection between the vertical wall panel, the horizontal joint and the foundations;
• Diagonal cracks have not been observed in the panels;
• The strains in the uniformly distributed horizontal and vertical reinforcement point to relatively low stresses in it, which means that it remains non-activated in the panels, hence this reinforcement can be reduced to 0.1%;
• The hysteretic relationships show that the elements have a relatively high ductile capacity, while the displacements are relatively small. This is considered favourable for the structural response to actual seismic effects. The non-pinched hysteretic relationships point that the model has high capacity of energy dissipation.

Prefabricated large panel reinforced concrete systems as well as other reinforced concrete systems exposed to dynamic seismic effects of moderate and high intensity suffer nonlinear deformations of constituent elements whereby their initial strength and rigidity is considerably decreased.

A special section in Bozinovski (1997) is dedicated to this problem, where a detail review of more important hysteretic models used for dynamic analysis of structures is presented. In Figure 7 an original hysteretic model is proposed. Some of more important conclusions given at the end of this section are as follows:

Figure 7. Proposed original hysteretic model.

• The nonlinear dynamic response of structures exposed to gravity and seismic loads is of a special interest, especially for construction of buildings in seismically active regions.
• Considering the vertical wall panels and vertical and horizontal joints which can be used for modelling their strength rigidity and deformability under cyclic loads, polygonal diagrams for the strength - deformability relationships are proposed. Through the polygonal strength-deformability diagram the proposed mathematical relationship gives the possibility to include all phases of the mechanism of behaviour during the formation and the opening of the cracks, principal reinforcement yielding, concrete crushing in the compressed part of the joint with much or less sliding along the length of the horizontal or vertical joints up to complete failure. In this way the vertical walls panels, the vertical and horizontal joints, i.e. all the elements along the height of the building influencing the behaviour of the structure are included in definition of the structure dynamic response.
• With such a mathematical model and computer program it is possible to define the response of the model of the large panel structure under expected seismic effect of different intensity and frequency content.

In the final section, Conclusion and Recommendations, presented in Bozinovski (1997), some general conclusions are pointed out. Two of them are as follows:

• The analysis of the behaviour of prefabricated large panel systems in seismically active regions, after earthquakes occurrences, shows that they can behave favourably under seismic effects if designed in compliance with the requirements for aseismic design of these systems and their elements;
• The joints are favourable elements for dissipation of considerable part of the seismic energy which is obvious from the width of the hysteretic strength-deformability diagrams, especially in joints loaded also with gravity loads although designed against occurrence of considerable sliding;

This chapter is based on the material received with the kind assistance of Prof. Eisenberg (1998), chairman of the Russian National Committee for Earthquake Engineering. It points out the necessity of using new materials as well as new structure of some precast elements due to the requirements mentioned in the last changes of the Code for construction in Russia.

The existing equipment of the industrial construction in Russia includes 380 plants with nominal production 34.3 mln m² production per year. The real production in 1994 was 12.4mln m² that means 36% of the nominal one. Equipment mass in all these plants reaches the value of 1.5mln tonnes. This is the main issue, which shows that any change of the general purposes of the factories for precast constructions and elements can cause significant losses as well as problems in satisfying the country necessities with residentials. In this connection it has to be confessed that in spite the decreased production volume of precast residentials, this type of construction is going to be still leading one in the near future as well as increasing the participation of the lower buildings.

The new edition of the building codes in Russia, No SNIP II-3-79 (introduced by MINSTROY) acting now is connected with radical "hardening" of the thermal insulation characteristics of the existing buildings. During the first stage (up to 2000 year) the required resistance of the thermal insulation fencing constructions has to increase twice and during the second stage it has to reach 350% of the initial one. This is the reason for a critical review of the traditionally used materials and constructions.

Masonry structures in Russia are traditionally used because of aesthetic, exploitation and economic reasons. Sudden energy price increasing makes necessary a new discussion on masonry as traditional material in terms of processing and technical characteristics. To satisfy the technical requirements according to the Russian building Code new edition, mentioned above, masonry walls have to become double in thickness, which is absolutely unacceptable from economic point of view. To realise the second stage in increasing thermal insulation characteristics new effective materials and layered structure details have to be introduced.

All these comments are valid also for outer walls made from light concrete (density 900kg/m³). As an exception cellular concrete blocks, widely used abroad, can be considered. The preliminary calculations and already constructed residential show that timber beam or outer walls are also uneconomic for this second stage. In this way it is drawn out that most suitable specially for residential of small height are layered fencing walls.

These new requirements related mainly to thermal insulation reflect to changes into different construction systems - large panel system, cast in situ and masonry system and combined system.

• Large panel system is used in a very huge variety with different structural schemes. The thickness of the horizontal diaphragms is 220mm or 160mm. Inner walls are constructed compact, made by light or heavy concrete. World practice and a lot of studies in this field show that large panel system can meet the requirements of modern civil construction practice considering architecture aspects, city and construction planning, building exploitation and environment prevention. Some new solutions related to horizontal diaphragm supports and better external view of these buildings has been proposed by Russian specialists in Moscow.
• Cast in situ and mixed, cast in situ and precast, system. The most important here it is the possibility that industrial outputs are used for the concrete preparation. Such concrete elements are applicable in any combination with precast elements, but it is not the usual case in Russia because of high price of the formwork, low temperatures in the most part of Russia, technological problems, etc. Russian practices show very good behaviour of precast structures with cast in situ inner walls. Layered walls here can be both bearing and not bearing with thickness of the bearing one 25-38cm. Bearing outer walls are prescribed to be used up to 5 storey buildings. Considering 9 storey and higher, the outer masonry walls can be just hanging.
• Combined system. Different elements and materials are simultaneously used in these structures - panel structures, small blocks, special masonry, etc. Rational architectural decisions as well as better application of the different material and structural details are realised this way. Such systems have started to be used in Russia in the recent time. The system is very good specially for 1 or 2 storeys structures as family houses of cottage type providing large spans without changing the formwork. Keeping the formwork is a great advantage from economical standing point.

As an extract from the studies, investigations and analysis in the different chapters the following conclusions can be made:

• The response of the prefabricated buildings during earthquakes, as well as the performed theoretical and experimental investigations on this type of buildings in some countries, show that being exposed to seismic excitations they are prospective and safe structures if the design and construction face the special requirements for building in seismic regions.
• The allowed shortcomings in the design and construction of the prefabricated buildings in some countries are related mainly to shortcomings in the urban planning, not enough good architectural appearance of the buildings and their layout decision, as well as to deficiency in insuring of thermal, sound and hydro insulation. Most of the allowed shortcomings are not characteristic of the prefabricating methods in the construction but are result of different subjective and objective reasons, that have given negative reflection also on the traditional construction methods. For example, in Bulgaria due to these reasons are also included the following ones: trend for satisfying of the keen necessity for dwellings but neglecting other needs for complex urban planning; monopoly situation and wish of the producer to simplify the work by production of as less as possible types and marks of prefabricated elements that leads to monotony in the architectural decision of the buildings; lack of competition between the construction enterprises; weak construction control and so on.
• Good behaviour during seismic excitations is shown by the buildings constructed using different large panel systems. That is due to the favourable outline of the scheme of the bearing structure formed by the bearing wall panels and the considerable absorption of the seismic energy in the connections as well as the well designed joints between the elements.
• The adequate design and construction of the connections are of decisive importance for the reliability of the prefabricated buildings in seismic regions. For this reason in the national standard documents as well as in the documents elaborated for the European Union, special attention is paid to this problem. It concerns mainly the increased safety in their design (increased values of the safety coefficients) and considering other requirements related to the location of the connections or to the way of their design in order to ensure adequate ductility. Thus for the frame reinforced concrete structures it is recommended not to make the connections at the joint beam-column itself, for the large panel buildings it is recommended the connections to be constructed in a manner to be able to absorb a great part of the seismic energy. The investigations carried out in this report show that special attention must be paid to the insuring of the reliability of the connections to corrosion too. This is valid for the existing buildings also and especially for observation of the process of carbonatization of the concrete. Details of the measures that should be undertaken in this respect for the construction of new connections as well as for insuring the already constructed ones, are given in part 2.2 of this report.
• In comparison with the traditional methods of construction, the prefabrication methods have been applied for relatively short period of time. That is why the further development and improvement of the prefabrication methods should be accompanied by carrying out of additional intensive theoretical and experimental investigations. That is specially valid for the design and construction of prefabricated buildings in seismic regions. In this respect the international cooperation plays an important part. This cooperation should be extended and accelerated and the attention must be drawn to some more effective forms of exchange of experience. For example, to hold more often a special international seminar with the participation of more countries, it could be an intensified form of exchange of experience in the field of applying of prefabricated buildings in seismic regions.