POST-EARTHQUAKE REPAIR AND STRENGTHENING OF REINFORCED CONCRETE BEAM-COLUMN CONNECTIONS ( THEORETICAL & EXPERIMENTAL INVESTIGATION )

In this study the effectiveness of all the repair and strengthening techniques proposed by the United Nations Industrial Development Organization (UNIDO) Manual and by Eurocode 8: part 1-4 for reinforced concrete beam-column joints damaged by strong earthquakes is investigated experimentally and analytically. Five one-half-scale exterior beam-column joint specimens were submitted to reverse cyclic pseudo-static displacements. Three of these specimens were then repaired by the epoxy pressure injection technique or by the removal and replacement technique. The other two specimens were strengthened by partial three-sided jacketing. All the repaired and strengthened specimens were then subjected to the same displacement history as that imposed on the original specimens. It can be concluded that all the repair and strengthening techniques proved to be satisfactory.


INTRODUCTION
Investigations into recent earthquake damage in Greece (Thessaloniki, 1978;Corinth, 1981;Kalamata, 1986;Aegion, 1995;Athens, 1999) have shown that, in many cases, damaged areas of reinforced concrete buildings were localised in beam-column connections.Furthermore, considering the commonly accepted idea that failure of joints may quickly lead to general failure (Park and Paulay, 1975), the important issue of the effective repair or strengthening of beam-column connections damaged in cyclic loading has arisen (Popov and Bertero 1975, Rodriguez and Park 1991, Penelis and Kappos 1997, Elnashai 1997).
In trying to address the issue of the effectiveness of any repair and strengthening technique, it is essential to bear in mind that the behaviour of reinforced concrete connections involves the influence of complex interacting phenomena such as shear, bond, confinement, fatigue, which even independently are not yet well understood.Therefore, since unanswered questions exist even concerning the behaviour of undamaged connections designed to withstand seismic excitations, it is justifiable to allow a certain measure of uncertainty for the mechanics of applied repair and strengthening techniques.Additional questions concerning the effectiveness of generally applied reinforced concrete joint repair and strengthening techniques also arise from the existing variety of shear reinforcement design practices.

REPAIRANDSTRENGTHENINGTECHNIQUESFOR BEAM-COLUMN JOINTS ACCORDING TO UNIDO 1983 AND ACCORDING TO EUROCODE 8: PART 1-4 (STRENGTHENING AND REPAIR OF BUILDINGS) 1995
In 1983, the United Nations Industrial Development Organization, with the participation of several countries in the Balkan region, produced a manual based on experience gained in this region, which provides mainly qualitative guidelines for the repair and strengthening of buildings.Some case studies are also presented in this manual (UNIDO (1983), Rodriguez and Park (1991)).
Eurocode 8: part 1-4 (strengthening and repair of buildings) ( 1995) covers the repair and strengthening of buildings and, where applicable, monuments, considering commonly used structural materials (concrete, steel, masonry and timber).
Field reports after damaging earthquakes often indicate that beam-column joints are one of the most vulnerable structural regions.Under earthquake loading, joints often suffer shear and/or bond (anchorage) failures.Three possible repair and strengthening techniques exist, namely:

(i) Epoxy injections
Epoxy injections can be applied in the repair of damaged joints with slight to moderate cracks without damaged concrete or bent, or failed reinforcement.
(ii) Local replacement of damaged concrete and steel ("equal section method") Removal and replacement should be carried out for heavily damaged joints with crushed concrete, buckled longitudinal bars or ruptured ties.Depending on the amount of concrete removed some additional ties or reinforcement may be added.Before concreting, the existing joint should be saturated with water as necessary (UNIDO (1983), EC8: part 1-4 (1995)).
(iii) Reinforced concrete jacketing In the case of heavily damaged joints of space frames, a reinforced concrete jacket is required, which can be located in the joint area only.The reinforced concrete jacketing of a joint is performed in such a way that all the members connected at the joint collaborate together.

TESTS CONDUCTED
An investigation was conducted at the University of Thessaloniki to evaluate the effectiveness of the techniques proposed by UNIDO 1983 and the Eurocode 8: part 1-4 (Strengthening and repair of buildings) 1995 for the repair and strengthening of reinforced concrete beam to column connections damaged by severe earthquakes.More specifically, five reinforced concrete exterior beam-column subassemblages were constructed.The subassemblages were subjected to cyclic lateral load histories so as to provide the equivalent of severe earthquake damage.The damaged specimens were then repaired or strengthened according to the UNIDO Manual Techniques (UNIDO 1983) and according to Eurocode 8: part 1-4 (1995).These repaired and strengthened specimens were again subjected to the same cyclic lateral load history.'Ire rmwred resp:.mehistories of the original and repaired or strengthened specimens were subsequently compared and evaluated.

BACKGROUND AND PREVIOUS RESEARCH ON THE PERFORMANCE OF REPAIRED AND/OR STRENGTHENED RIC BEAM• COLUMN SUBASSEMBLAGES
Despite the many unanswered questions related to the behaviour of reinforced concrete structures repaired by epoxy injection, this technique has been used extensively in the aftermath of past earthquakes around the world.Popov and Bertero (1975), presented a comparison of the performance of a reinforced concrete interior beam-column subassembly tested under cyclic loading with its performance after repairing it with epoxy resin.They observed that the bond around the reinforcing bars in the joint region once destroyed does not seem to be completely restored by epoxy injection.UNIDO 1983, EC8: part 1-4 (1995) and NEHRP 1985 (FEMA-97) state also that epoxy injection is not effective in restoring the bond between reinforcement and concrete.Corazao et al. (1988) investigated the effectiveness of different repair and strengthening techniques in restoring or improving the properties of reinforced concrete beamcolumn subassemblages damaged by earthquake-type loading.Based on their test results, they concluded that large variations in the performance of subassemblages repaired by epoxy injection can occur depending on the quality of the injection work.Data obtained from a testing subassemblage after repairs involving the removal and replacement of the damaged concrete in the beam and beam-column joint region demonstrated that the stiffness and strength of the specimen had been completely restored.Lee et al. (1980) investigated the effectiveness of repair of reinforced concrete exterior beam-column subassemblages.
The epoxy injection technique and the removal and replacement technique using high-early-strength materials were used to repair the beams of the subassemblages.They observed that because of the increase in beam strength due to the use of high strength repair materials, there is the possibility of damage moving from the beam to the unrepaired joint and column.The same was also observed by Corazao et al. (1988).
The effectiveness of the epoxy injection technique for the repair of reinforced concrete beam-column connections damaged by cyclic loading was also investigated by French et al. (1990), Karayannis et al. (1998).They concluded that the epoxy injection technique is an effective method to repair earthquake damage of beam-column joints.They also found out that bond between reinforcement and concrete in the joint region was restored by this repair procedure.Gulkan (1977), Corazao et al. (1988), Alcocer and Jirsa (1990), Paultre and Mitchel (1990), Mitchel (1995), Tsonos (1999) showed that beam-column joint specimens strengthened by jacketing exhibited higher strength, greater stiffness and better energy dissipation capacity than the asbuilt specimens.However, Gulkan (1977) and Corazao et al. (1988) concluded that the beam-column joint is undoubtedly the most difficult structural member to strengthen because of the great number of elements, connected in this region.
Alcocer and Jirsa (1990) and Tsonos (1999) assessed experimentally the behaviour of beam-column subassemblages strengthened by jacketing where the joint was confined with collar stirrups.They found that this joint reinforcement provides adequate confinement and shear capacity to the joint.
An important work in this field was recently published by Hakuto et al. (2000).Test results from poorly detailed reinforced concrete beam-column joints retrofitted by jacketing with new reinforced concrete are reported in this work.They concluded that: "The jacketing of beam-column joints with new reinforced concrete was identified as a useful technique for enhancing the stiffness, strength and ductility of poorly detailed as-built beam-column joint regions.The technique, however, is very labour-intensive and the placement of the new joint core hoops, passing through holes to be drilled in the existing beams is difficult".

EXPERIMENTAL SET -UP LOADING SEQUENCE
Each specimen was tested before and after repair or strengthening, under reverse cyclic loading in the Laboratory of Reinforced Concrete Structures at the Aristotle University of Thessaloniki.The general arrangement of the experimental set-up is shown in Figure l(a).All specimens (before and after repair and/or strengthening) were subjected to a large number of cycles applied by slowly displacing the beam's free end, according to the load history shown in Figure l(b), without reaching the actuator stoke limit.The amplitudes of the peaks in the displacement history were 15 mm, 20 mm, 25 mm, 30 mm, 35 mm, 40 mm, 45 mm, 50 mm, 55 mm, 60 mm and 65 mm, corresponding to drift angle R of up to 6.5%.One loading cycle was performed at each displacement amplitude.An axial load approximately equal to 045Pb (Pb: balanced column load) was applied to the columns of the subassemblages and kept constant throughout the test.Approximately 10 electrical-resistance strain gauges were bonded in the reinforcing bars of each specimen of the programme.u------------    As seen in Figure 2(c), the joint transverse reinforcement of specimens F 1, F2 and L1, did not satisfy the requirements of the ACI-ASCE Committee: sh= 70 mm> Sh(required) = 200/4 = 50 mm (A,h = A,h(required) = 90 mm 2 ).The values of the joint shear stress factor were greater than 1.0 for both the specimens F1 and L1, whereas the values of flexural strength ratio were less than 1.40 for both the specimens F2 and L1, see Fig. 2(c).Thus, the beam-column connections of the original specimens can be expected to fail in shear.

Failure mode of the original subassemblages
Specimens A 1 and Ei: Failure mode of specimens A1 and E1, as expected, involved the formation of a plastic hinge in the beam near the column juncture.The formation of a plastic hinge caused severe cracking of the concrete near the fixed beam end of each subassemblage.The behaviour of the original specimens A1 and E1 was as expected and as documented in the seismic design philosophy of the modem codes (ACI-318 (1995), EC2 (1993), EC8 (1994)).
Significant inelastic deformations occurred in the beams' longitudinal reinforcement in both specimens (strains of over 40.000µi: were obtained in the beams' longitudinal bars), while the shear mechanisms of their joints remained in the elastic range.Figure 4 shows the strain gauges data of hoop reinforcement of the joint regions of both subassemblages A1 and E1• As is clearly shown in Figure 4, the maximum strain recorded in the joint hoop reinforcement was below 2500µ1:.
One difference between the failure modes of specimens A1 and E1 was that hairline cracks appeared in the joint region of E1, and partial loss of the concrete cover in the rear face of the joint of E1 took place during the three last cycles of loading (9'1\ 10 th and 11 th ), while the joint region of subassemblage A1 was intact at the conclusion of the test (see Fig. 3).

Specimens F1 and L1:
The specimens F1 and L1 failed by yielding of the joint ties during their first cycles of loading.A shear failure occurred, as expected, in the joint region of both the original structures F1 and L1• This shear failure occurred before the formation of plastic hinges in their beams.The maximum strain recorded in the longitudinal bars of the beams was below 2500µ£.The longitudinal column reinforcement of specimen F1, consisting of 014 bars, was not buckled in the joint region.On the contrary, the longitudinal column reinforcement of specimen L1, consisting of 010 bars, was significantly buckled in the joint region (see Fig. 3).The beams in both specimens F1 and L1 remained intact at the conclusion of the tests (Fig. 3 ).
Specimen F2: Specimen F2 developed a flexural hinge in its beam.Damage occurred both in this region and on the beamcolumn joint (Fig. 3 ).
Strains of over 40.000µE were obtained in the beam longitudinal bars of F2.Yielding of the joint ties of F2 was recorded after the 5 th cycle of loading.

UNIDO and EC8: Part 1-4 repair and strengthening techniques, specimens RAi, REi, RF1, RF2 and RL1
The repair procedure applied to specimens A1 and E1 included the following operations: Removal and replacement of the crushed and loose concrete in the beam near the fixed end of both specimens A 1 and E1 by a premixed, non-shrink, rheoplastic, flowable and non-segregating mortar of high strength with 9.5 mm maximum size aggregate.
Removal and replacement of the spalling and loose concrete cover of the rear face of the joint of specimen E1 by a thick layer of epoxy resin paste.
Superficial sealing of all visible cracks with a thick layer of epoxy resin paste except for plastic inserts located along the cracks which serve as ports allowing inlet access for thin epoxy resin to be injected into the system and outlet access for air to escape from the crack voids.Injections under pressure of thin epoxy resin into the crack system of the damaged area of the joints until total fillup.The whole infusion procedure requires special care in order to avoid local air entrapment.
After the hardening period of the high-strength mortar used for the replacement of crushed concrete of the beams of both subassemblages A1 and Ei, the resin injection procedure was also applied to restore the crack system of the damaged area of both specimens around the added high-strength mortar.
The repaired specimens remained unloaded during the period of resin hardening (for at least seven days).
The highest percentage of crushed concrete in the joint region of specimen F 1 was lost during the tests, as is clearly shown in Figure 3.The longitudinal column reinforcement of this specimen, consisting of 014 bars, was not buckled in the joint regions.
Repair was performed as follows: The remaining concrete in the joint regions and the damaged concrete cover of part of the columns' critical regions was removed with a chipping hammer.The concrete surfaces were cleaned of dirt and dust.The same pre-mixed mortar of high strength with 9.5 mm maximum aggregate size, which was used for the repair of the beams of specimens A1 and E1, was also used for the replacement of the crushed concrete in the joint region and of the loose eover of the columns' critieal regions of RF1 (Fig. 5(a)).The values of the flexural strength ratio of RF1 were higher than 1.40 and the values of joint shear stress were lower than 1.0 ..jf'{ MPa, (Specimens A1 (RA 1), E1 (RE 1) and F 1 (RF 1) are first reported in earlier publieations (Tsonos (2001(Tsonos ( , 2002))) and are reported here again in order to present all the repair and strengthening techniques for beam-column joints according to UNIDO (1983) and according to Eurocode 8: part 1-4 (1995)).
Although, it is strongly recommended in the UNIDO Manual that columns and beam-column joints be jacketed on all four sides for the optimum performance in future earthquakes, it also gives examples of three-sided or two-sided jackets of columns and beam-column joints.These types of jackets are inevitable when there are adjacent structures abutting the original building to be strengthened from one or more sides.Thus, it was considered worthwhile to investigate the seismic performance of exterior reinforced concrete subassemblages upgraded by three-sided jackets.It is worth noting that the strengthened beam-column joint subassemblages in the literature were all four-sided jackets.
The removal and replacement of crushed and loose concrete in the beam near the fixed end of specimen F2, by the aforementioned mortar of high strength was carried out.The beam of specimen L1 was intact at the conclusion of the tests (Fig. 5

(b)).
A strengthening technique has been applied to the subassemblages, consisting of both raising the reinforcement quantity to minimum levels and covering the whole joint area and parts of the critical regions of the columns of the specimens with a three-sided cement grout jacket reinforced with additional ties (Fig. 5(b)).The additional longitudinal reinforcement was placed at each comer of the jacket, which was then welded to the existing column reinforcement.The existing longitudinal column reinforcement of specimens F2 and L1 consisting of 010 bars was significantly buckled in the joint region (Fig. 3).Thus, additional column reinforcement was needed to replace the buckled reinforcement.This additional longitudinal column reinforcement was also needed in order to increase the flexural strength ratios of specimens F2 and L1, and especially that of specimen L1 (MR(F2) = 1.30< 1 .40,MR(L1) = 0.72 << 1.40, Fig. 2).The flexural strength ratios of strengthened specimens RF2 and RL1 were significantly higher than those of the original specimens F2 and L1, respectively (Fig. 5(b)).
The same pre-mixed mortar of high strength with 9.5 mm maximum size of aggregate was used for the construction of the cement grout jacket of specimens RF2 and RL1.
As shown in Fig. 5 The compressive strengths of the high strength mortar used for the removal and replacement of crushed concrete of specimen RF 1 and for the construction of the jackets of specimens RF2 and RL1 were 70 MPa, 68 MPa and 66 MPa, respectively.Electrical-resistance strain gauges were bonded in the reinforcing bars within the joint region of both the original and repaired and/or strengthened subassemblages.

Additional joint transverse reinforcement
Two additional horizontal ties were placed in the joint region of specimens RF1, RF2 and RL1 in order to increase their shear strength (Fig. 5(c).
The values of the flexural strength ratio were higher than 1.40 and those of the joint shear stress were lower than 1.0 ,J"f{' MP a for all the repaired and strengthened specimens RA1, RE1, RF1, RF2 and RL1 (Figures 2 and 5).The joint transverse reinforcement of specimens RF1, RF2 and RL1 with the two additional ties was 08 at 35 mm centres.This reinforcement satisfied the requirements of the Committee (ACI 352R-85): sh = 50 mm = 200 rnm/4 (A,1i = A,h(required) = 90 mm 2 ).
The provision of transverse reinforcement, made of short bars placed and connected under the bends of a group of rebars, was made to ensure the anchorage of the beam bars in the joint region (Eurocode 8 -1994, see Fig. 5).

SpecimenE1
Specimen F1 Specimen F2 Specimen L 1   The strengthened subassemblages could, therefore, be expected to fail in flexure and, more specifically, to develop flexural hinges in the beams without severe damage concentration in the joint regions.

BEHAVIOUR OF REPAIRED AND STRENGTHENED SPECIMENS
Failure modes of the repaired and/or strengthened subassemblages RAi, REi, RFi, RF2 and RL1 Both specimens RA1 and RE1 repaired by epoxy injections exhibited similar failure modes, nearly identical to that of the original specimen A1.Thus, the failure mode of both the repaired specimens RA1 and RE1 involved the formation of a plastic hinge in the beam near the column juncture and damage concentration in this region only, see Figure 6.In both specimens the rupture of some longitudinal beam reinforcing bars in the plastic hinge region took place during their last two cycles of loading (10 th and 11 th for RE1, 8 th and 9 th for RA1).It is obvious that the failure mode of the repaired specimen RE1 was better than that of the original specimen E1.
Failure mode of specimens RF1, RF2 and RL1, as expected, also involved the formation of a plastic hinge in the beam near the column juncture and damage concentration in this region only.During the final cycles of loading of specimens RF1, RF2 and RL1 when large displacements were imposed, the damaged concrete cover could not provide adequate support for the beam longitudinal reinforcement.As a result, buckling of the beam reinforcement in specimens RF 1, RF2 and RL1 occurred after the 6'\ 6tl 1 and 9 th cycles of loading, respectively.
Strains of over 40.000µi: were obtained in the beam longitudinal bars of all the strengthened specimens RA1, RE1, RF1, RF2 and RL1.The maximum strain recorded in the joint hoop reinforcement of all the repaired and strengthened specimens was below 2500µ1:.In Figure 7 are demonstrated strain gauge data of the hoop reinforcement of the joint regions of specimens RAi, RE1 and RF1.
The three-sided jacketing of beam-column joints is more critical than the four-sided jacketing, especially in the rear face of the joint along the column, where the hooked ends of the beam longitudinal reinforcement move outward to split the cover.As is clearly demonstrated in Figure 6, the rear faces of both specimens RF2 and RL1 were intact at the conclusion of the tests.It is worth noting that the rear face of specimen RF1 repaired by the removal and replacement technique was also intact at the conclusion of the tests (see Fig. 6).
The beam in the subassemblage RF2 was repaired using the removal and replacement technique and was much stronger than the original one.It is worth noting that the increase in beam strength due to the use of high strength repair mortar did not result in damage moving from the beam to the strengthened joint and column of specimen RF2• As is clearly shown in Figure 6, both the column and the joint of specimen RF2 were intact at the end of the tests.

Load drift angle curves
The performance of the test specimens is presented herein and discussed in terms of applied shear-versus-drift angle relations.Drift angle R, which is plotted in the figures that follow, is defined as the beam tip displacement D. divided by the beam half span L, and is expressed as a percentage (see inset on Fig. 8).Plots of applied shear-versus-drift angle for all the specimens A1, RA1, E1, RE1, F1, RF1, F2, RF2, L1 and RL1 are shown in Figure 8.
The original beam-column specimens A1, E1, F1, F2, and L1 showed stable hysteretic behaviour up to drift angle R ratios of 4.0 percent, 4.0 percent, 2.0 percent, 4.0 percent and 2.0 percent respectively (Fig. 8).They showed a considerable loss of strength, stiffness and unstable degrading hysteresis beyond drift angle R ratios of 4.5 percent, 2.0 percent, 4.5 percent and 2.0 percent respectively (Fig. 8).
The repaired specimens RA1 and RE1 exhibited stable hysteresis up to 6 th cycle of drift angle R, of 4.5 percent, after which a significant loss of strength began due to noticeable buckling of the beam reinforcement (Fig. 8(a)).The extreme loss of strength, stiffness and energy dissipation capacity observed in specimen RA 1 during 8 th and 9 th cycle of loading, and in specimen RE1 during the 11 th cycle of loading, was due to the fracture of almost half of the beam reinforcing bars during these cycles of loading.
The repaired specimen RF 1 and the strengthened specimen RF2 exhibited stable hysteresis up to the 6 th cycle of drift angle R, of 4.0 percent, after which a significant loss of strength began, due to the noticeable buckling of the beam reinforcement (Fig. 8(b)).
The strengthened specimen RL1 showed stable hysteresis up to the 9 th cycle of drift angle R of 5.5 percent, after which also a serious loss of strength began, due to the remarkable buckling of the beam reinforcement (Fig. 8(b)).
Comparison of the strength, stiffness and energy dissipation capacity between the original Ai, L1 and repaired and/or strengthened RA1, RL1 subassemblages respectively.
For a further evaluation of the effectiveness of the UNIDO 1983 and EC8: part 1-4 (1995) repair and strengthening techniques in restoring and increasing the strength, stiffness and energy dissipation capacity of the damaged subassemblies, it is interesting to compare the peak-to-peak stiffness, the energy dissipated and the peak strength observed for every load cycle of the original specimens A 1 and L1 with those of the repaired and strengthened specimens RA1 andRL1- The peak-to-peak stiffness and energy dissipated for every load cycle of each specimen A1, L1, RA1 and RL1 are illustrated in Fig. 9(b) and Fig. 9(c), respectively.-~ i i i i iii I II II _111_ "' -20 ~ -20
From these diagrams, it is also clear that the strengthened specimen RL1 achieved significantly increased strength, stiffness and energy dissipation capacities as compared with those of the original specimen L1, even in the large displacement amplitude cycles of drift angle R between 5.5 percent and 6.5 percent.It is now necessary to establish a relationship between the average normal compressive stress er and the average shear stress,.From Equations (1) and ( 2):

THEORETICAL CONSIDERATIONS
( It is well known that V jh h C where a is the joint aspect ratio.Thus, CT= a' T (5) The maximum principal stresses are given by Mohr's circle (Fig. lO(c)) and the following expressions can be recovered: From the diagram of the behaviour of concrete under biaxial stresses (Kupfer et al 1969), it was found that the branch AB could be represented by a 5th degree parabola (Tegos 1984) (Fig. lO(d)).Thus, for branch AB: (7) where fc is the increased joint concrete compressive strength due to confining, which is given by the model of Scott, Park and Priestley (1982) according to the equation .,::Also, f ~ is the concrete compressive strength, and K is the parameter of the model (Scott et al 1982) and is expressed as: where p, is the volume ratio of transverse hoop reinforcement, and fyh is the yield strength of the transverse reinforcement.
Substituting Eqs. ( 5) and ( 6) into Eq.( 7) and using -c = y .Jf: gives the following expression: Assume here that and cq, 10) is then transformed into: The solution of the system of equations ( 11)-( 13) gives the beam-column joint ultimate strength.This system is solved each time for a given value of the joint aspect ratio using standard mathematical analysis.
For simplicity's sake, the presentation of the above methodology was for exterior beam-column joints.The approach is the same for interior joints.For the development of the above formulation, it has been assumed that bond conditions of the beam and column bars anchoring or passing through the joint region are generally favourable.
The proposed shear strength formulation can be used to predict the failure mode of the subassemblages and thus the actual values of connection shear stress.Therefore, when the computed joint shear stress is greater or equal to the joint ultimate capacity Ycal :C: Yui,, the predicted actual value of connection shear stress will be near Yutt because the connection fails earlier than the beam(s).When the calculated joint shear stress is lower than the connection ultimate strength Ycal < Yui,, then the predicted actual value of the connection shear stress will be near Yeab because the connection permits its adjacent beam(s) to yield.
More details concerning the above formulation can be found in Tsonos (1996Tsonos ( , 1997Tsonos ( , 1999)), where the validity of the 91 formulation was checked using test data for 40 exterior and interior beam-column subassemblages that were tested in the Structural Engineering Laboratory at the Aristotle University in Thessaloniki, as well as data from similar experiments carried out in the United States.
The improved retention of strength in the beam-column subassemblages, as the values of the ratio Yea1 / Yui, decrease, was also demonstrated.It is worth noting that for Yea!/ YuJt ::;; 0.50 the beam-column joints of the subassemblages performed excellently during the tests and they remained intact at the conclusion of the tests (Tsonos (1996(Tsonos ( , 1999))).
The shear capacities of the repaired and/or strengthened beam-column connections of specimens RA 1, RE1, RF1, RF2 and RL, were computed using the above methodology.

CONCLUSIONS
The following conclusions are drawn based on the work presented herein.
1.The original strength, stiffness and energy dissipation capacities were restored in the subassemblages repaired by epoxy resins.The bond between reinforcement and concrete also appeared to be restored by the repair procedure.In general, the epoxy-repaired cracks did not reopen in the tests of the repaired structures, new cracks tended to develop in the concrete adjacent to the repaired cracks.
2. The beam in both subassemblages RA1 and RE1 were repaired using epoxy injection technique, as well as the removal and replacement technique and were stronger than the original ones.The increase in beam strength due to the use of high strength repair materials did not result in damage moving from the beam to the unrepaired joint and column of specimen RA1 (as observed by Lee et al. 1980).Both the column and the joint of specimen RA 1 were intact at the conclusion of the tests.

Specimen RF1 repaired by the Removal and Replacement
Technique and specimens RF2 and RL1 strengthened by three-sided jacketing exhibited significantly increased strength, stiffness and energy dissipation capacities as compared with those of original specimens F1, F2 and Li, respectively.For Yea!< Yui, , Ypred = Yea!(an overstrength factor a= 1.25 for the beam steel is included in the computations of joint shear stress Teal= Yea! ,jf;; MPa).

4.
All the repaired and strengthened specimens RAi, RE1, RF i, RF2 and RL1 developed flexural hinges in their beams near their column juncture.They showed high strength, without any appreciable deterioration after reaching their maximum capacity.The beam-column joints of all the repaired and strengthened specimens were intact at the conclusion of the tests.
i (a).Test setup ( dimensions in mm)

Figure 1 .
Figure 1.Test setup and loading sequence.
(b) both specimens RF2 and RL1 had the same three-sided cement grout jacket plus 010 longitudinal bars at each comer of the column connected by 08 supplementary tics at 70 mm centres.All longitudinal bars in the jackets extended into the beam-column region of the subassemblages.Both the original and repaired and/or strengthened subassemblages were constructed using deformed reinforcement.All (both the original and the repaired or strengthened) specimens' steel yield stress can be summarised as follows: 08 = 520 MPa, 010 = 535 MPa, 012 = 530 MPa and 014 = 520 MPa.(Note: 08, 010, 012 and 014 = bar with diameter 8 mm, 10 mm, 12 mm and 14 mm).

Figure 4 .
Figure 4. Applied shear-versus strain in beam-column joint hoop reinforcement of the subassemblages A 1 and E 1 •

Figure 5 .
Figure 5. (a) Details of repaired specimen RF1 (dimensions in cm), (b) Jacketing of beam-column connections of subassemblages RF2 and RL1 (dimensions in cm) and (c) Additional joint ties of specimens RFi, RF1 and RL1.

Figure 9 (
Figure9(a) compares the peak strength observed throughout the tests.The comparison is made by observing the ratio of the peak strengths of the repaired and strengthened subassemblages RA 1 and RL1, to that of the original subassemblages A1 and L1.

Figure 8 (
Figure 8(b ).Hysteresis loops of specimens F b RF b F 2, RF 2, L 1 and RL 1 • Figure I O(a) shows a reinforced concrete exterior beamcolumn joint for a moment resisting frame.Park and Paulay ( 1975) on the basis of experimental findings postulated the existence of two shear mechanisms in RIC beam-column joints.Thus, according to the approach of Park and Paulay, the total shear within the joint core is carried partly by a diagonal concrete strut, formed between the comers of the joint subjected to compression [see Fig. lO(a)], and partly by an idealized truss consisting of horizontal hoops, intermediate column bars and inclined concrete struts between shear cracks (Fig. IO(a)).Both mechanisms depend on the core concrete strength.Thus, the ultimate concrete strength of the joint core under compression/tension controls the ultimate strength of the connection.After failure of the concrete, strength in the joint is limited by the gradual crushing along the cross -diagonal cracks and especially along the potential failure planes (Fig.IO(a)).

Figure 9 12 ~Figure 9 Figure 10 .
Figure 9(a).Strength ratio of repaired or strengthened model to original models.

5 .
The proposed Techniques by the UNIDO and by Eurocode 8: part 1-4: Epoxy Injections, Removal and Replacement, and Reinforced Concrete Jacketing for the repair and strengthening of reinforced concrete beam-column joints have proven to be effective methods of repairing severe earthquake damage to this structural element.NOTATIONMR= sum of the flexural capacity of columns to that of beam 0 the parameter y at ultimate capacity of the connection [ y ult -K]