Effect of Intermediate Stiffeners on the Behaviors of Partially Concrete Encased Steel Beams

Partially concrete encased steel (PE) beams are composite steel beams and concrete elements that present several advantages, such as higher fire resistant, higher flexural capacity, and higher lateral torsional buckling resistant compared to bare steel beams. ,is paper reports an experimental study of eight PE beams under cyclic loading. ,e effectiveness of intermediate stiffeners, such as midspan stiffener and plastic hinge zone stiffener, in enhancing composite action and ductility of the PE beams was studied. ,e ductility performance of PE beams using strengthened beam-to-column connection and weakened beam-to-column connection was also investigated.,e test results show that the plastic hinge zone stiffener performed well and has the potential to replace shear connectors. Strengthened and weakened beam-to-column connections can be implemented in PE beams to enhance the ductility of the PE beams; with the details of both strengthened and weakened beam-to-column connections determined by bare steel shape instead of the whole section. In addition, the suggestions to prevent premature failure of weakened beam-to-column connection were provide.


Introduction
Fire-proofing material is usually used to provide fire resistance to the steel members in high-rise buildings.e fireproofing material may cause pollution to the environment; therefore, extra construction material is often needed to cover the fire-proofing material.Fully concrete encased steel beam or steel-reinforced concrete (SRC) beam, as shown in Figure 1(a), uses concrete to provide fire resistance to the steel in the beam; therefore, the pollution problem of fire proofing material can be avoided.However, the concrete brings extra vertical and seismic loads to the structure and the construction is comparatively more difficult and requires intensive labor.
e partially concrete encased steel beam shown in Figure 1(b) uses concrete in cooperation with fire-proofing coating to provide fire resistance to the steel beams.e concrete encasement on partially concrete encased beam can effectively prevent the temperature of steel from rising sharply [2][3][4].Compared to the SRC beams, this type of partially concrete encased steel beam, defined as PE beam hereafter, reduces the amount of concrete used, which reduces the vertical and seismic load on the structure.In addition, PE beams reduce the requirement for formwork, shoring, and steel cage fabrication.is reduces construction difficulties and cost.Furthermore, the presence of concrete infill between the steel flanges also contributes to PE beam to have higher bending resistance [5][6][7][8] and higher lateral-torsional buckling (LTB) resistance than bare steel beam [1,9].ese advantages have led to wide application of PE beam to be used in modern construction.e PE beam shown in Figure 1(b) is the type of beam investigated here.
Several studies have been conducted to examine the composite action on PE beams.Kindmann et al. [5] investigated the composite behavior of PE beams under monotonic loading.It was found that shear connectors were not necessary to develop composite action of the beams.However, the beams tested used tension reinforcement, which reduced the requirement for shear connectors.Nardin and Debs [6] examined the composite action of PE beams by testing simply supported rectangular beams with monotonic loading.e beam tested had enlarged bottom ange, which leads to higher requirement for shear connectors.Test results suggested that shear connectors are required for the beams to develop a composite action, and shear connectors installed on the bottom ange developed higher composite action than those installed on the web did.In addition, the slip between concrete and steel at the end of the test beams went beyond 2.0 mm at the time when the exural strength was developed.Hegger and Goralski [7] evaluated the exural capacity of PE beams which were integrated with concrete slab under both sagging and hogging moment.e test results showed that, when the beam under sagging moment, the exural capacity of the beam with and without shear connection nearly had no di erence.e loaddeformation diagram of beam with and without shear connection could be recognized without any di erences, with the load di erences about 1%. e measured slip at the end of the concrete encasement was always less than 0.2 mm. e shear force was transferred by friction forces between the anges of the steel pro le and the concrete encasement.While under hogging moment, the beam without su cient shear connection showed lower exural strength.Chen et al. [8] conducted monotonic and cyclic tests on PE beams with span length of 0.75 m and 1.5 m (shear span ratio of 1.5 and 3, respectively).e plastic moment capacity is exceeded for all the test specimens and also showed good ductility under both monotonic and cyclic loading.
At present, research on PE beam used shear connection to develop the composite action between the steel shape and the concrete encasement which was tested by simply supporting with monotonic loading [5][6][7].However, when a PE beam forms part of a moment resisting frame, as shown in Figure 2, the columns may provide constraint to the beam at the beam ends and suppress the slip between concrete and steel shape surfaces, as schematically explained in Figure 2(b).
is column constraint should have positive e ect on the development of beam composite action.However, when the distance between beam ends, or columns, is too far, the e ect of the column constraint may be reduced.In this case, extra sti eners installed between the length of PE beam, referred as intermediate sti eners, may provide extra constraint and further enhance the composite action.For the beam shown in Figure 2 subjected to earthquake type loading, a sti ener at the midspan, as shown in Figure 2(c), where bending moment changes direction, may be helpful.In addition, a sti ener installed near the plastic hinge zone, where the requirement for composite action is the highest in the beam span, should be advantageous to PE beam to develop its exural strength.
In steel structures, it is common practice to move the plastic hinge away from the column surface to prevent premature fractures of the welds connecting the beam ange and the column ange.Strengthened beam-to-column connections (SBC) and weakened steel beam-to-column connections (WBC) have been suggested to solve this problem [10][11][12][13][14].A PE beam has the same concerns as a steel beam has.And, it is of interest to the authors how PE beams perform when SBC or WBC is used.
In this study, a total of 8 PE beam specimens were fabricated and tested under cyclic loading.e e ectiveness of column constraint and intermediate sti eners, such as midspan sti ener and plastic hinge zone sti ener, was investigated.In addition, the suitability of SBC and WBC to PE beams is assessed.

Test Specimens.
e designations of the 8 specimens tested are shown in Table 1.Each test specimen was mounted to a pair of columns, as shown in Figure 3. e test setup con guration used in this research is similar with the test scheme conducted by Sudibyo and Chen [15].e columns and the end-plate connections possessed very high sti ness and strength so that their deformation can be neglected.All the specimens were symmetrical with respect to their center lines and had a clear span length of 1200 mm. e bottom end of each column was connected to a rigid base through a hinge.
Figure 4(a) shows the geometry and dimensions of the cross section used.
e cross-sectional steel shape used, which was H 150 × 75 × 5 × 7, had a ange width-to-  Advances in Civil Engineering thickness ratio of 5.4.e ends of cross-sectional steel shapes were connected to the end plates with complete penetration welds.ASTM A36 steel with a nominal yield stress of 250 MPa and concrete with a speci ed compressive strength of 30 MPa were used to design the specimens.ere were 4 specimens in the PW series.In this group, a type of weakened beam-to-column connections proposed by Chen et al. [11], as shown in Figure 5(a), were used.eoretically, the geometry of the ange tapering should be determined according to the exural strength of the whole PE section (i.e., steel shape plus concrete).However, it becomes tedious.Instead, for the purpose of simplicity, the geometry of the ange tapering, as shown in Figure 5(a), was determined based on steel shape alone.Figure 6(a) shows the moment capacity distribution along the beam, where M pd (24.5 kN•m) is the design plastic moment (also the nominal design exural strength) of H 150 × 75 × 5 × 7 calculated with nominal yield stress of the steel.e plastic moment is calculated as speci ed in Table 4.1 of ANZI/AISC 360-16 [16].Projecting the moment capacity of Section C to the beam end resulted in a beam end moment of 0.92 M pd .
When determining the geometry of ange tapering, the beam end moment was controlled in between 0.90 and 0.95 M pd and the distance between the critical section (Section C) and the end of the welding access hole was kept 25 mm.
After the determination of the geometry of the ange tapering, the distribution of nominal design moment capacity M nd , based on whole section, along the beam was calculated and is shown in Figure 6(b).e nominal moment capacity is the moment capacity of the whole composite PE section, with complete composite action, based on the plastic stress distribution method as speci ed in Section I1.2a of ANZI/AISC 360-16 [16] and schematically explained in Figure 4(b).e factor β 1 used in the calculation is referred to ACI [17].e M nd (26.9 kN•m) is calculated with nominal yield stress of the steel.e ideal condition for weakened beam-to-column connection is that the sections in the tapered ange region reach their design moment capacity simultaneously [11].It can be found from Figure 6(b) that the sections between C and D reach their moment capacity almost at the same time.is shows that determining the geometry of the ange tapering based on steel shape alone  should be quite satisfactory.e critical section located at Section C. Projecting the design moment capacity of Section C to the beam end resulted in a usable design beam end moment of 0.93M nd .ere were 4 specimens in the PS series.In this group, strengthened beam-to-column connections, as shown in Figure 5(b), were used.Similar to PW specimens, the cover plates of PS specimens were designed based on steel shape alone for simplicity.
e length of the cover plates was determined in such a way that the critical section located at a distance of 25 mm away from the end of the welding access hole.In order to prevent failure at beam end, the moment capacity of the combination of anges and cover plates was kept at least 1.2 time of M pd .Figure 7(a) shows the design plastic moment distribution along the beam.Figure 7(b) shows the design nominal moment capacity, based on whole section, along the beam.e critical section located at C and the usable beam end moment was 1.09M nd .e H shape detail used in this research is the same with the steel conguration conducted by Sudibyo and Chen [15]; therefore, the moment capacity distribution of the steel shape, Figures 6(a) and 7(a), are the same as those in [15].e specimens with "SC" in their designation were equipped with shear connectors, as shown in Figure 8(a).e specimens with "MS" in their designations had a midspan sti ener installed, as shown in Figure 8(b).e specimens with "HS" in their designations had one plastic hinge zone sti ener installed at each end of the beam.e plastic hinge zone sti eners were installed at the position 175 mm, which is 1.2 times the beam depth, from the end plate, as shown in  e mechanical properties of the H steel shape are listed in Table 2.All the specimens used ready-mixed concrete without coarse aggregate.
e specimens were cast in a horizontal position in two stages: the concrete on one side of the steel beam was cast rst, and the concrete on the other side of the steel beam was cast a week later.
e compressive Advances in Civil Engineering strength on each side of the steel beam was 38.1 and 38.8 MPa, respectively.e average value (38.4 MPa) was taken as the concrete compressive strength for the specimens.

Test Setup and Loading
Procedure.e test setup is shown in Figure 9.A servo-controlled hydraulic actuator with a capacity of 500 kN was connected to the right column.
e actuator was equipped with a built-in load cell and built-in linear variable di erential transducers (LVDT) to measure, respectively, the applied force P and the lateral displacement at the load point.e lateral displacement at the load point is de ned as displacement of the right column.e displacement of the left column was measured by an LVDT at the beam center.e average of displacement of the left and right column is de ned as the lateral displacement Δ. Rotation gauges R1 and R2 were installed on the end plates of the beam, and rotation gauges R3 and R4 were installed   Advances in Civil Engineering 175 mm from each beam end.e beam segments between R1 and R3 and between R2 and R4 are defined as plastic hinge zones.e axial link was used to reduce axial force in the specimens.
A quasi-static load was applied under displacement control according to the lateral displacement history shown in Figure 10.Δ y is defined as the measured lateral displacement of the frame when the moment at beam ends, as indicated in Figure 11, reached nominal moment capacity M na (29.56 kN•m) of the test beams.M na is the moment capacity of the whole composite section, calculated based on the actual material strength, complete composite action, plastic stress distribution method (Section I1. 2a ANZI/ AISC 360-16) [16] as shown in Figure 4(b).A Δ y of 7.5 mm was obtained and used for all the specimens.It is noted that the moment distribution presented in Figure 11 is based on the assumption of antisymmetrical moment distribution of the beam.Since the test scheme (Figure 3) and the H shape configuration (Figure 5) of the specimens used in this research are the same as those used by Sudibyo and Chen [15], the moment distribution of the beam (Figure 11) is the same as that used by Sudibyo and Chen [15].e load test ended when the strength of the specimen deteriorated more than 20%.

General Behavior.
e P versus Δ hysteretic loops for the specimens are shown in Figure 12, and the test results are summarized in Table 3. e drift angle α of the frame is defined as the corresponding Δ divided by L c .e maximum loads in the positive and negative directions are designated, respectively, as P + peak and P − peak .e maximum load for the tested specimen P exp is defined as the average of P + peak and P − peak .SBC and WBC are intended to prevent failure of the welds connecting the beam and the column, and the failure of the beam was designed to occur at the critical section of the beam, as shown in Figures 5-7.erefore, the following analysis is performed based on beam's moment capacity in the critical section.
According to the moment distribution shown in Figure 11, the moment at the critical section corresponding to P exp , designated as M exp , can be calculated as follows: It is worth noting that M exp is an underestimate since the strength of the beam at both ends was, very likely, not developed at the same time.e nominal moment capacities of the critical section (M na ) c for PS specimens and PW specimens were 29.6 kN•m and 25.1 kN•m, respectively, as listed in Table 4. e strength ratio c is defined as the ratio of M exp and (M na ) c .e plastic moment of the steel shape in Advances in Civil Engineering the critical section, calculated using actual stress of steel, (M p ) c for PS specimens and PW specimens were 26.7 kN•m and 22.6 kN•m, respectively, as listed in Table 4. And, the strength ratio c ′ is defined as the ratio of M exp and (M p ) c .
e strength ratios c and c ′ of all the specimens are listed in Table 3.It is noted that the ratio (M na ) c /(M p ) c was 1.11 for both PS and PW specimens.is indicates that the existence of the concrete raises the moment capacity of the steel beam by about 11%.
e rotation of the left and right plastic hinge zones of the beam are designated as θ L and θ R , respectively.θ L was calculated by subtracting R3 readings from R1 readings, and θ R was calculated by subtracting R4 readings from R2 readings.Figure 13 shows the lateral load versus plastic hinge rotation hysteresis loops for specimen PS-HS.e P vs. θ skeleton curves based on first cycle of each drift ratio excursion for PS and PW series are shown in Figures 14  and 15, respectively.e yield rotation θ y and ultimate rotation θ u were determined based on P vs. θ skeleton curve of each specimen.Figure 16 shows positive P versus θ L skeleton curve based on the first cycle of each drift angle excursion.e ultimate rotation θ + uL is defined as the rotation corresponding to 85% of P + peak in the descending portion.e elastic rotation θ + yL , as indicated in Figure 16, is the estimated elastic rotation corresponding to P + peak .e plastic hinge rotation capacity θ + pL was the difference between θ + uL and θ + yL .rough the procedure aforementioned, θ + pL , θ − pL , θ + pR , and θ − pR for each specimen were determined.e plastic hinge rotation capacity, θ p , of each specimen is defined as the average of θ + pL , θ − pL , θ + pR , and θ − pR .

Behaviors of PS Series.
e failure procedure for all PS specimens was similar.Concrete crushing and steel flange buckling, as shown in Figure 17(a), was observed when the applied load reached P + peak and P − peak .After that, cracks at the end of the cover plate started to develop and eventually caused the fracture of the steel flange, as shown in Figure 17(b).e hysteresis loops of the specimens are quite stable up to one or two cycles after local buckling of steel flange occurs.
Figures 18(a)-18(d) show the condition and the crack pattern of PS specimens after the first 5% cycle.e damage on the right end of the specimen is more severe than that on the left end, which indicates that the strength of the beam at the two ends did not develop at the same time.And, P exp and M exp underestimate the real strength of the beams.Consequently, specimen PS-SC, which is considered possessing complete composite action, has a strength ratio c (0.99) slightly less than 1.0.8 Advances in Civil Engineering

Behaviors of PW Series.
Due to the tapering of the flange, concrete occupied the space where flange was tapered, indicated as "flange concrete" in Figure 19.When the flange concrete was subjected to compressive force, it was pushed sideward and caused concrete damage to occur earlier than that of PS specimens, as shown in Figure 20.Since the flanges of PW specimen were tapered and a smaller width-tothickness ratio is then resulted, flange local buckling occurred somewhat later than PS specimens did, as indicated in Figure 12.
e crack patterns of PW specimens, as shown in Fig- ures 18(e)-18(h), also show that the concrete damage on the right end of the specimen is more severe than on the left end.PW specimens experienced more severe concrete damage than the corresponding PS specimens.
Cracking of the steel flange was observed after peak loads were reached.Cracking, as indicated in Figure 21(a),  Advances in Civil Engineering started from the center of the flange just above the end of the welding access hole.e crack then extended toward the tapered region of the flange.Figure 21(b) shows the flange fracture of PW-NB specimen.e premature flange fracture and the extra concrete damage near the flange tapering region reduced the experimental strength of the specimens.As a result, the strength ratio of PW-SC (0.92) is even lower than that of PS-SC (0.99) by about 7%.
e premature flange fracture should be deferred, and higher flexural strength and ductility should be developed if larger access hole-taper start clearance is provided [11].

Effectiveness of Intermediate Stiffeners.
Since premature failure of the PW specimens was observed, it is more reliable to compare the effectiveness of composite action based on test results from PS specimens.Take PS-SC as a reference, the P exp /(P exp ) SC ratio and θ p /(θ p ) SC ratio of PS-NB, PS-MS, and PS-HS are calculated and listed in Table 5, where (P exp ) SC and (θ p ) SC are the P exp and θ p of PS-SC specimen.
e P exp /(P exp ) SC ratio for PS-NB (0.94) is apparently lower than 1.0. is indicates that column constraint alone is not quite enough to develop full composite action of the PE beam regardless of the fact that plastic hinge rotation capacity of PS-NB is no less than that of PS-SC.
e P exp /(P exp ) SC and θ p /(θ p ) SC ratios of PS-MS are 0.97 and 1.09, respectively.ey are higher than those for PS-NB specimen.is indicates that the midspan stiffener is able to provide extra constraint and develop higher degree of composite action of the PE beam.However, the effectiveness may diminish for longer beams.
e P exp /(P exp ) SC for PS-HS is 0.99 which is in the range of experimental error.In addition, the θ p /(θ p ) SC is as high as 1.17.Test data indicate that the plastic hinge zone stiffener is able to develop a very high degree of composite action and to enhance significantly the ductility of the PE beams.Usually, the length of the plastic hinge zone is in the range of 1.0 to 1.5 times the beam depth, no matter how long the beam is.
erefore, the position of the end stiffener does not vary so much, and the effectiveness of the plastic hinge zone stiffener will not diminish as the beam length becomes larger.Test results indicated that the plastic hinge zone stiffener used in this study has the potential to replace shear connectors on developing the beam composite action.With the existence of plastic hinge zone stiffener in the PE beam, less shear connectors or shear studs are required to hold the concrete encasement between the beam's webs; therefore it will simplify the fabrication of PE beam.

Strengthened and Weakened Steel Beam-to-Column
Connection.
e θ p that PS specimens achieved ranges from 5.21 to 6.11% which is much greater than 4%.
is is considered sufficient for seismic design of highly ductile members as required by AISC seismic provision [18].For PW specimens, although premature failure occurred, the θ p achieved (4.52 to 4.75%) is also sufficient for beams in highly seismic zones.Test results shows that the application of SBC and WBC on the PE beams worked satisfactorily.In addition, designing SBC and WBC based on moment capacity of the H steel shape instead of the whole section seems to have worked well for the specimens tested in this study.
e average P exp of PS specimens (69.3 kN) is 25% higher than that of PW specimens (55.5 kN). is phenomenon can be attributed to three factors.e first factor comes from the beam-to-column connection scheme itself.e usable design beam end moment for SBC is 1.09 M n , which is 17% higher than that of WBC (0.93 M n ).e influence of this factor will be reduced as the moment gradient of the beam becomes smaller.e second factor is the premature failure of the flange.And, the third factor is the concrete damage next to the tapered flange.10 Advances in Civil Engineering e averaged θ p value of PS specimens is 20% higher than that of PW specimens.e premature fracture of the flange and the damage of the concrete next to the tapered flange make major contributions to this phenomenon.Providing sufficient "access hole-taper start clearance," as shown in Figure 22, should be able to avoid premature failure of the flange.
For the PW specimens tested here, a "flange net width w n " can be estimated according to the fracture path observed, as shown in Figure 23, and Section B4.3b of AISC specification [16], as follows: e calculated w n equals to 73 mm which is smaller than the flange gross width (75 mm). is provides an explanation for the premature failure of the flange.It is suggested that, to prevent the premature failure of WBC, an access hole-taper start clearance that ensures a net flange width w n greater than flange gross width should be provided.
e concrete damage next to the tapered flange should can be mitigated by removing the concrete in the tapered flange area (termed as flange concrete herein), as the shaded region indicated in Figure 19. is can be accomplished by installing thin steel plates in the area of tapered flange before concrete is casted, as indicated in Figure 19(a).ick plate welding is considered to have adverse effect on ductility development of strengthened beam-tocolumn connections.Since the PS specimens tested here were small-scale specimens, the thick plate welding effect did not show here.However, in the real structure, the thick plate welding effect is likely to emerge, and the ductility of the strengthened connection may be reduced.

Conclusions
Eight partially concrete encased steel (PE) beams were fabricated and tested under earthquake type loading.e effectiveness of midspan stiffeners and plastic hinge zone stiffeners in enhancing composite action and ductility of the PE beams was studied.
e strength and ductility of strengthened beam-to-column connection and weakened   12 Advances in Civil Engineering beam-to-column connection were also investigated.Based on the experimental results reported herein, the following conclusions can be drawn: (1) e plastic hinge zone sti ener used in this study can successfully develop the composite action of PE beams.e exural strength of the beam can be fully developed and the ductility of the beam was 17% higher than that of beams with shear connectors.e plastic hinge zone sti ener has the potential of replacing shear connectors for developing composite action of the beam.(2) e midspan sti ener used in this study provides certain e ect on developing composite action of PE beams.However, the composite action developed was less complete as the beams use shear connectors.(3) e strengthened and weakened beam-to-column connections, which were used to enhance the ductility of steel beams, can also be implemented in PE beams.e ductility of the specimens tested is high enough to be used in buildings in highly seismic zones.(4) Design of strengthened and weakened beam-tocolumn connection based on the cross-sectional steel shape alone, instead of the whole section, was able to lead to high ductility of the beams.is can greatly simplify the design of strengthened and weakened beam-to-column connections of PE beams.(5) Premature ange fracture may happen if the access hole-taper start clearance is not long enough.It is suggested that an access hole-taper start clearance that ensures a net ange width greater than ange gross width should be provided.(6) For beams with weakened beam-to-column connections, the concrete near the tapered ange region was subjected to stress concentration and su ered more severe damage.is brings more concrete cracks, lower exural strength, and low exural ductility to the beams.
is phenomenon can be mitigated by removing the concrete next to the ange at the ange level.

Symbols c ′ :
Distance from the top of concrete to the neutral axis of the composite section f y : Yield stress of steel f u : Ultimate stress of steel L b : Length of the beam (refer Figure 11) L c : Length between the column's pinned connection to the beam center (refer Figure 11) L cs : Distance between the two critical section (refer Figure 11) L s : Distance between the columns (refer Figure 11)  Lateral displacement Δ y : Measured lateral displacement of the frame when the moment at beam ends reached M na .

aFigure 2 :
Figure 2: PE Beam in a moment resisting frame.(a) Moment distribution of the beam under earthquake type loading.(b) Column constraint at beam end.(c) Plastic hinge zone sti ener and midspan sti ener.

Figures 5
Figures 5 and 8(b).Specimens with "NB" in their designations used neither shear connector nor sti ener.e mechanical properties of the H steel shape are listed in Table2.All the specimens used ready-mixed concrete

Figure 7 :
Figure 7: Design moment capacity distribution of the beam with SBC connection.(a) Steel shape; adapted from [15].(b) Whole section.

Figure 17 :Figure 18 :
Figure 17: Damage of selected PS specimens.(a) Local buckling of the steel.(b) Fracture of the steel flange of specimen PS-NB.

Table 1 :
Details of test specimens.

Table 3 :
Test result of test specimens.

Table 4 :
Moment capacity of the section.
Calculated with the material's nominal strength.(b)Calculatedwith the material's actual strength (test value).

Table 5 :
e effectiveness of concrete constraint schemes.
Moment at the critical section corresponding to P exp M na :Nominal moment capacity is the moment capacity of whole composite section, based on actual material strength (M na ) c : Nominal moment capacity of critical section M nd :Design nominal flexural strength of the whole section (composite section), based on nominal material strength M pd :Design plastic moment of the steel shape, based on nominal material strength (M p ) c : Plastic moment of the steel shape at critical section, based on actual material strength P: Strength ratio M exp /(M na ) c c ′ :Strength ratio M exp /(M p ) c