You are on page 1of 26

EFFECTS OF MULTIPLE DAMAGE-HEAT STRAIGHTENING REPAIRS ON STEEL BEAMS

Keith Kowalkowski
Staff Engineer
Ruby and Associates, P.C.
E-Mail: kowalkowski@rubyusa.com

Amit H. Varma
Assistant Professor
School of Civil Engineering
Purdue University
E-Mail: ahvarma@purdue.edu

Purdue University
School of Civil Engineering
550 Stadium Mall Drive
West Lafayette, IN 47907
Phone: (765) 496-3419
Dept. Fax: (765) 496-1105

Submission Date: January 19, 2007
Word count: Text + Tables + Figures = 5000 + 1000 + 1500 = 7500

1

ABSTRACT
Experimental investigations were conducted to evaluate the effects of multiple cycles of damage followed by heat
straightening repair on the structural properties and fracture toughness of steel bridge beams. Six W24x76 (or
equivalent) beam specimens made from A7, A588, or A36 steel (two beam specimens of each steel type) were
subjected to three cycles of damage followed by heat straightening repair. In each cycle, the beam specimens were
damaged statically in weak axis bending, and repaired using several half-depth Vee heats and an external restraining
force. Material specimens were machined from the flanges of the damaged-repaired beam specimens and tested
according to ASTM standards. The material test results indicated that multiple damage-repair cycles do not have a
significant influence on the elastic modulus, ultimate stress, and surface hardness of the steels. Multiple damage-
repair cycles generally increase the yield stress and reduce the ductility of the steels. The charpy V-notch (CVN)
fracture toughness of A7 and A36 steel beams repaired with maximum heating temperature (Tmax) equal to 650°C
was lower than the fracture toughness of the corresponding undamaged steels. The CVN fracture toughness of A588
steel beams repaired with Tmax equal to 650°C, and the A36 steel beam repaired with Tmax equal to 760°C was much
higher than the fracture toughness of the corresponding undamaged steels. Material specimens that were removed
from locations closer to the flange-web interface consistently had lower fracture toughness and ductility.

Keywords: heat treatment, steel bridges, notch toughness, stress, strain, hardness, rehabilitation.

2

In some cases. Avent et al. Avent et al. Combinations of Vee. recommendations for heat straightening patterns. or spot heats) and locations to reverse the plastic deformations and repair the damage. These guidelines are used by some state departments of transportations (DOTs) across the U. (7. T (torsional).INTRODUCTION Over-height trucks occasionally (impact) collide with steel girders of highway bridges. 4). and local buckling or denting at the point of impact. 6). The results indicated that the damage-repair cycles increase the yield and ultimate stress. which includes categories S (strong axis). Several state highway agencies use heat straightening to repair damaged steel girders. 7.50Mp) and Vee heats in the damaged regions (flanges). This paper presents the results of a research project focusing on this topic. and repaired using restraining moments equal to 50% of the plastic moment (0. It consists of applying heating in specific patterns (Vee. and spot heats are used to repair the damaged composite steel girders. strip. Researchers have conducted experimental investigations to determine the deformations and residual stresses produced by heat straightening (4. and spot heats are used to repair the local buckles and dents (2. and some do not have any heat repair guidelines (1). BACKGROUND Significant research has been conducted on the heat-straightening repair of steel girder bridges. Tension coupons were taken from damaged-repaired regions near the apex. The beams were damaged by weak axis bending (Category W). 11. and 8 damage- repair cycles. and hardness) of bridge steels have also been studied by several researchers (7. 4. Several heating cycles are applied until the residual deformations reach within tolerance levels and the damage is repaired. The composite steel girders typically undergo categories T and L structural damage due to collision. 2. maximum heating surface temperatures. 3). The structural damage caused by the collision consists of permanent plastic deformations and can be described using the damage characterization scheme developed by Avent and Mukai (2. but significant after 4-8 damage-repair cycles. 3. These guidelines include: techniques for measuring damage. This further reinforced the recommendation to limit steel bridge girders to two damage-repair cycles. Heat straightening is a cost-effective technique for repairing structural damage caused by collisions. where Vee heats are used to repair the out-of-plane deformations of the flange. 10) Currently.S. (7) suggested a limit of two damage-repair cycles for steel girders. W (weak axis). strip. Composite steel girder bridges subjected to this type of collision damage in Michigan are most frequently made from A7 or A373 steel (1). (7) also investigated the effects of multiple damage-heat straightening repair cycles on the structural properties of A36 steel W6x9 beams.. DOTs have significant interest in developing rational guidelines for evaluating the acceptability of steel girders subjected to multiple damage-repair events (1). 3). 6. Avent and Fadous (10) subjected a composite W24x76 A36 steel beam to four damage-repair cycles. strip heats are used to repair the web. and decreases the ductility of repaired steels. 8). The results from these studies indicate that heat straightening increases the yield and ultimate stress. while maintaining the steel surface temperature below 650°C (1200°F) for mild steels and 595°C (1100°F) for quenched and tempered steels (2. and the effects of damage and repair on steel beams (6. The effects of heat straightening on the structural properties (yield and ultimate stress. This raises concerns regarding their condition and serviceability after multiple damage-repair events. This includes out-of-plane deformation of the bottom flange. The results from these studies were used to develop empirical equations and guidelines for conducting heat straightening in the field (2). while other DOTs use in- house research. and decrease the ductility of the steel. a crack formed on the convex side of the web yield line. heat-straightening repair guidelines have been developed by Shanafelt and Horn (9) and Avent and Mukai (2). and open end of the Vee. and L (local buckling) damage for steel shapes. Heating is applied using oxygen-fuel torches. ductility. Based on this research. 5. Hence. This recommendation is based on research conduced by Avent et al. They subjected the steel beam specimens to 1. center. The reduction in ductility was moderate after 1-2 cycles. limits on restraining forces during repair. Significant research has been conducted on the heat-straightening repair of damaged steel shapes (4. and tolerances for residual deformations after repair. 12). 3 . and tested according to ASTM standards. the same steel bridge girder is subjected to multiple damage and heat straightening repair events over its life. During the fourth cycle. Current guidelines indicate that the same steel girder should not be subjected to more than two damage-heat straightening repairs. causing significant structural damage. 8). 5.

The damage and repair parameters included were: (i) the inelastic damage strain (!d=30. Both specimens were subjected to three cycles of damage-repair with !d equal to 40. The experimental results and major findings from the phase-1 investigations are presented in detail elsewhere (13. and A36 steel) were tested. 0. 0. and A36-Beam 2 was repaired with Tmax equal to 760oC (1400°F). and 60!y.25Mp-y for cycles 1. and A588-Beam-2 was repaired with Mr equal to 0. 60. Mp-y is the weak-axis plastic moment capacity of the beam. while the remaining parameters were held relatively constant. The smaller damage strain (!d=30!y) was used because it was more detrimental than larger damage strains according to the phase-1 findings. A7-Beam-1 was subjected to the three cycles of damage-repair with !d equal to 30!!y. A36-Beam 1 repaired with Tmax equal to 650oC (1200°F). and 0. and A36 steel specimens were repaired with the maximum heating temperature (Tmax) limited to 650°C (1200°F). 4 . The parameter that was varied between the two beam specimens of each steel type was based on the phase-1 findings presented in (13. or 90 times the yield strain). A36. EXPERIMENTAL INVESTIGATIONS – PHASE 2 Test Matrix Table 1 shows the test matrix for the experimental investigations of phase-2.50Mp-y for all three damage-repair cycles. two specimens of each steel type (A7. Some A36 steel specimens were repaired with Tmax equal to 760°C (1400°F) or 870°C (1600°F). respectively.25Mp-y. respectively. and 3. This project consisted of two phases.50. A total of ninety steel plate specimens were tested. !d is the maximum damage strain. The first phase conducted laboratory-scale tests on steel plate specimens by subjecting them to multiple damage-heating repair cycles. The maximum heating temperature (Tmax) was equal to 650oC (1200°F) for all three cycles. and 3. 4. and to evaluate and verify the findings from phase-1. #p is the plastic displacement after damage. and A588) bridge steels. 0. This paper presents the experimental investigations and findings from phase-2 of the project. Tmax was equal to 650°C (1200°F) for all three cycles. As shown in the Table. one major damage- repair parameter was varied between the two specimens (of each steel type).25. A7-Beam-2 had to be repaired with Mr equal to 0. A588. Most of the A7. A36-Beam-2. • The heating temperature was the parameter varied between the two A36 steel beam specimens. A7-Beam-1. The nomenclature consists of the steel type followed by the specimen number. 20.25. 3.50. Both specimens were subjected to three cycles of damage-repair with !d equal to 30!y. !y is the undamaged steel yield strain. A588-Beam-1 was repaired with Mr equal to 0. In Table 1. and Nr is the number of multiple damage-repair cycles. etc. 2. A588. A7-Beam-2 was subjected to three cycles of damage-repair with !d equal to 90. Typically.RESEARCH PROJECT The authors (13) have conducted a research project focusing on the effects of multiple damage-heat straightening repair cycles on the structural properties (elastic modulus. respectively.50Mp-y because numerous Vee heats were required to repair the larger damage strains. The second phase conducted large-scale experimental investigations on steel beam specimens to further evaluate the effects of multiple damage-heat straightening repair cycles on the structural properties and fracture toughness of steel beams. or 5 cycles). and 20!!y for cycles 1. 14) for A7 and A36 steel. and needed further evaluation and verification in phase-2. The effects of each parameter were evaluated by varying it (within its range). 14): • The damage strain was the major parameter varied between the two A7 steel beam specimens. It includes the nomenclature for the beam specimens along with the details of damage and repair parameters. while holding all other parameters constant. and 3. ductility) and fracture toughness of common (ASTM A7. • The restraining moment was the parameter varied between the two A588 steel beam specimens. Tmax is the maximum temperature achieved during heating. 14). 2. for example. yield and ultimate stress. and (iii) the number of damage-repair cycles (Nr=1.25 Mp-y for cycles 1. Mr is the maximum restraining moment. 60. 2. All the specimens were subjected to three damage-repair cycles because that was the limit recommended by phase-1 (13. A7-Beam-1 was repaired with Mr equal to 0. and Mr equal to 0. (ii) the restraining stress ("r=25 or 50% of the yield stress). 2.

which is difficult to quantify in the field and duplicate in the laboratory with the facilities available. Additional details of the loading frames are presented in (13. in weak axis bending. Test Setup Figure 3 shows photographs of the test setup used to subject the beam specimens to damage (weak axis bending) followed by heat straightening repair. Figure 1(a-b) shows the damage for two of the three bridges. 15). which is also similar to the heat-straightening repair of the damaged flanges used by the SBC (2. thus providing enough data for confidence in the results. The authors documented each repair performed by the SBC and observed some overheating in the range of 690-760ºC (1300-1400°F) in some Vee heats. Several half- depth Vee heats were applied to the bottom flange using oxy-acetylene torches as shown in Figure 2(c). The flange plates were attached to the web plates using 10mm. The heat-straightening repair of beam flanges damaged in weak axis bending consists of several partial depth Vee heats spread along the length of flanges.Testing Approach The authors visited three heat straightening repair sites to determine the heat straightening techniques. Based on these similarities. Additional details of the test setups are presented elsewhere in (13. both flanges of the steel beam are subjected to identical damage and repair histories. The major limitation of the testing approach is that it does not simulate the impact damage of composite steel girders. on either side of the beam midspan. A portion was removed from the beam mid-length to fabricate material coupons for evaluating the undamaged A7 steel properties. fillet welds. This repair procedure is used most commonly by the SBC and is similar to the one used by Avent and Mukai (2. These Vee heats had a constant base width of 125 mm. The testing approach was designed to approximate the damage and heat straightening repair procedures used by the SBC. Additionally. the steel beam specimens were subjected to category-W static damage and repair for the experimental investigations. and subjected to concentrated forces at the mid-span. 5 . and procedures employed by the Michigan Department of Transportation (MDOT) Statewide Bridge Crew (SBC). the fascia girders of simply supported composite steel bridges were damaged by collision with over-height trucks and repaired by heat straightening. limitations. The flange plates were 230mm wide and 19mm thick. The actuator applied the damaging (upward) force by pushing the loading frame against the flanges of the beam specimen. The concentrated forces were applied using a double-acting hydraulic actuator that was bolted and post-tensioned to the laboratory strong floor. The A588 and A36 beam specimens were built-up sections with details and dimensions similar to the W24x76 A7 steel beam specimens.9m long A7 beams specimens were fabricated from either side of this central portion. The beam specimens were roller supported at the ends. The half- depth Vee heats were spaced along the damaged (plastic) regions as shown in Figure 2(d). A588. The restraining force was applied as two equal concentrated forces acting at 120 mm. and the web plates were 530mm wide and 13mm thick. Figure 4 shows a schematic of the loading frame used to apply concentrated forces at the beam midspan in both the upward and downward directions. It consisted of applying an external restraining force in the direction opposite to the impact force using the equipment shown in Figure 2(a). and A36 steels (two of each steel type). This testing approach facilitates much better control of the damage strain and the restraining stress in the beam specimens during damage-repair cycles. Strip heats were applied to the top of the web just below the concrete slab as shown in Figure 2(b). The composite steel girders exhibited large out-of-plane deformations of the bottom flange. resulting in Vee angles of 45-55°. 3). The various components of the test setup are identified in these Figures. Test Specimens Six beam specimens were fabricated from A7. which provides twice the material for obtaining samples for material testing. 15). 3). The heat-straightening repair procedure was also similar for all three cases.3m long) provided by MDOT. respectively. 3). Portions of the flange and web plates were retained for fabricating material coupons to determine the undamaged A588 and A36 steel material properties. The out-of-plane flange deformations of a composite steel girder (see Figure 1) are similar to the flange deformations of a steel beam damaged in (category W) weak axis bending (2. The A7 beam specimens were fabricated from a decommissioned W24x76 A7 steel beam (7. and two 2. Material specimens for tension and fracture toughness testing can be obtained from both flanges. In all three cases. Figures 3(a) and (b) show photographs of a beam specimen before and after damage.

Two pressure transducers were used to measure the actuator pressures and calculate the applied forces. Figure 3(a) includes some details of the instrumentation layout for the beam specimens. The portions of the flanges subjected to high residual stresses undergo yielding at lower loads. which was computed using a fiber model of the cross-section and the measured undamaged steel uniaxial stress-strain curves for the fibers.Instrumentation Several sensors were used to monitor the experimental behavior during damage and repair. The mid-span deflections measured at the target !d in the first damage cycle were used as the target displacements for the subsequent multiple damage cycles. Several heating cycles were required to repair each damaged specimen. The curvature ( $) corresponding to the moment (M) at each station was obtained from the section M. More Vee heats were applied at the locations with larger plastic strains. The midspan deflections were measured using two displacement transducers that were placed on the opposite flange plates of the beam specimens. Each heating cycle consisted of applying two simultaneous Vee heats directly across from each other on opposite flanges using two oxy-acetylene torches. C. They could not be used in subsequent damage cycles because the heat straightening repair process produced small out-of-plane distortions in the beam flanges. L1. Figure 6 shows the location.) identified in Figure 6. Figure 5 superimposes the Vee heat locations on the plastic strain distributions to illustrate that the Vee heats were located and distributed along the beam length to repair most of the material subjected to plastic strains (damage). which corresponds to a Vee angle of 53°.e. The quarter point deflections were also measured using two displacement transducers that were placed on either side of the beam midspan (halfway to supports). 4). The behavior shown in Figure 7 is typical and representative of the damage behavior of all beam specimens. EXPERIMENTAL BEHAVIOR Damage and Repair Behavior Figure 7 shows the measured force-displacement behavior for all three damage cycles of specimen A36-Beam-1. which results in reduced flexural stiffness. The inelastic force-displacement behavior of the beam and the corresponding plastic strain distributions were calculated using simple numerical models. The torch was then moved slowly in a serpentine path from the apex to the open end of the Vee. The beam length was discretized into several stations.5 mm (tolerance) of the displacement before damage. distribution. and R2. The surface temperature of the steel was monitored using a hand-held temperature-indicating device. The heat straightening repair was conducted by subjecting each beam specimen to several cycles of restraining force (producing moment Mr) and heating. L1. The displacements and end rotations were used to establish the completion of heat straightening repair. R1. a spot heat was applied initially at the apex of the Vee until the target temperature (Tmax) was reached. This Figure shows that after each damage-repair cycle. The station curvatures ( $) were used to compute the plastic strain distributions over the section and along the beam length.. The portions of the flanges with 6 . and the bending moment (M) at each station was computed using static force equilibrium. Figure 5 shows an example of the number of Vee heats required at the various locations to complete the repair of specimen A7-Beam-2. and nomenclature of multiple overlapping Vee heats that were used to repair the damaged specimens. L2.$ response. The end rotations of the beam specimens were measured using two rotation meters. the flexural stiffness of the beam decreases and it resistance (strength) increases for larger displacements. This behavior is influenced by the residual stresses produced by the damage and heat straightening repair processes (3. R1. Figure 5 shows an example of the plastic strain distribution calculated for specimen A7-Beam-2 subjected to the damage strain of 90!y in the first cycle. Three longitudinal strain gages were also bonded to the flanges of the beam specimen at the mid-span. etc. It shows the plastic strain distribution as a function of the distance from the midspan. In each Vee. i. These strain gages were used to monitor the maximum damage strain (!d) during the first damage cycle. Heat Straightening Procedure The Vee heat locations and distributions along the beam length were determined based on the plastic strain distributions produced by the damage process. Vee heats were applied until the midspan displacement of the beam specimen was within 1. The Vee heats were alternated between the various Vee locations (C. The depth and width of each Vee heat were approximately equal to 115 mm. The time-temperature (T-t) response of the heated steel was measured using a non-contact infrared thermometer.

Figure 8 shows the experimental behavior of specimen A7-Beam-1 during the third damage-heat straightening repair cycle. A588. Figure 10(a) shows the original location of these tension coupons in flange A.4°C (40°F) were used to compute the mean fracture toughness (FTo) and the 95% confidence interval of the fracture toughness for the undamaged steels.) were in accordance with (16). The results from the CVN impact tests at 4. ultimate stress ("uo). R1. As shown in Figure 9(a). "u/"uo.e. Each column of specimens was centered at one of the central Vee heats. As shown in Figure 9(b). at least 50 mm from the flange web interface) and fabricated from the mid-thickness of the flange plates. Y. The notches were located along the centerline of the Vee heats. The results from the uniaxial tension tests were used to compute the average elastic modulus (Eo). The displacement-time curves indicate that approximately 4-5 mm midspan deflection (repair) was achieved during each Vee heat with restraining moment (Mr) equal to 0. ultimate stress ("u). 17. yield stress ("y). 7 . L1. and eo). The uniaxial tension coupons and the CVN specimens were fabricated from the two flanges of the damaged-repaired beam specimens. yield stress ("yo). and the normalized values of E/Eo. Twelve CVN specimens were fabricated from the central region of flange B. Figure 8 focuses on the experimental behavior measured during the first four Vee heats. Y. "uo. restraining forces. !y/"yo. The behavior shown in Figure 8 is typical and representative of the behavior of all beam specimens during heat straightening repair. and (c) Rockwell hardness tests conducted according to ASTM standards (16. 18). These results were normalized with respect to the corresponding undamaged steel material properties (Eo.significant residual strains undergo strain hardening immediately after yielding due to the prior damage-repair excursions and history of the material. and ductility (% elongation. It includes the measured surface temperatures. The number and locations of the uniaxial tension coupons and CVN specimens were similar to those shown in Figure 9. and ductility (% elongation eo) for the undamaged steels. The surface temperature-time curves indicate that the steel surface heats rapidly (in less than one minute) to the target value (Tmax) and cools slowly (over 15 minutes) to room temperature before the next Vee heat is applied. L1 and R2. e) for the damaged-repaired steels. which were identified as flanges A and B. Rockwell hardness tests were conducted on the surfaces of CVN specimen C-2 to compute the average surface hardness (HRBo). and A36 steels are summarized in Table 2 Material Test Results The uniaxial tension test results for coupons (X. These results for the undamaged A7.50Mp-y. and oriented to measure the through thickness fracture toughness. the CVN specimens were located within 65 mm of the flange edge (i. The Rockwell hardness tests were conducted on the surfaces of CVN specimens C-2. The nomenclature for identifying the CVN specimens consisted of the Vee heat location (C. X. and innermost coupon closest the flange-web junction. etc. and A36 steel beams. thickness. and the quarter and midspan displacements as functions of time. center. where 1 is closest to the flange edge and 4 is closest to the flange-web interface. "yo. It also identifies the nomenclature used to identify the tension coupons. Figure 9(b) shows the original locations of these CVN specimens in the flange. or R1) followed by the numbers 1 to 4. Three tension coupons were fabricated from the central Vee heated region within 65mm of the edge of flange A. The applied restraining force increases slightly as the steel undergoes thermal expansion due to heating. Standard (ASTM) tests were also conducted on material samples obtained from the flange plates (A and B) of undamaged A7. respectively. Material Testing Procedures Standard (ASTM) tests were conducted on material specimens fabricated from the flanges of the beam specimens subjected to multiple damage-repair cycles. These tension coupons had a gage length of 50mm. A588. As shown. The applied restraining force decreases significantly as the steel undergoes yielding and then contraction due to cooling. the twelve CVN specimens were removed in three columns of four specimens each. The normalized surface hardness values (HRB/HRBo) for the damaged-repaired beam specimens are also included in Table 3. These included: (a) uniaxial tension tests (b) charpy V-notch (CVN) fracture toughness tests.. respectively. All the CVN fracture toughness tests were conducted after cooling the CVN specimens to 4. For clarity of presentation.4°C (40°F). and Z represent the outermost. and their remaining dimensions (width. which were applied sequentially to the Vee locations C. which is the temperature recommended for Zone 2 fracture toughness tests in (19). and e/eo are summarized in Table 3 for the six beam specimens. and Z) from flange A of the damaged-repaired beam specimens were used to calculate the elastic modulus (E).

These normalized fracture toughness values eliminate any bias produced by the value (high or low) of the undamaged steel fracture toughness. 5. and were lowest for the CVN specimen closest to the flange-web interface. and indicate the effects of damage-repair cycles more clearly. The results in Table 5 indicate that the fracture toughness values were very high for both A588 steel specimens. and the average (Avg-V) of the fracture toughness values for four CVN specimens in the same vertical column. The fracture toughness test results for the CVN specimens from flange B of the damaged-repaired A7. and they were consistently lower for the coupons (Z) closer to the flange-web interface. The CVN specimens close to the flange edge had very high toughness values (approximately 107-284% of the undamaged steel toughness). 8 . the increase in the yield stress was higher for the tension coupons closer to the flange-web interface. and 6. The toughness values were lowest at the center (C) Vee heat and reduced further for CVN specimens closer to the flange-web interface. ultimate stress. The damage-repair cycles reduce the ductility of A588 steel specimens to about 79-103% of the undamaged steel ductility. The tables also include the average (Avg-H) of the fracture toughness values for three CVN specimens at the same horizontal level. The toughness values were highest for the CVN specimens closest to the flange edge. The ductility values were slightly higher for the coupons (X) closer to the flange edge. which was subjected to larger damage strains. EXPERIMENTAL RESULTS Evaluation of Material Test Results for A7 Steel The normalized material properties in Table 3 indicate that the damage-repair cycles increase the yield stress of A7 steel to approximately 107-125% of the undamaged steel yield stress.4°C (40°F) are highlighted in bold. each table also presents the normalized fracture toughness values for the CVN specimens of both beam specimens. The normalized values are obtained by dividing the measured fracture toughness values by the mean fracture toughness (FTo) of the undamaged steel from Table 2. These results indicate that damage strain is an important parameter for A7 steel specimens. respectively. The results in Table 4 indicate that the fracture toughness values for specimen A7-Beam-1 were quite low. The overall average fracture toughness of A7-Beam-2 was 105% of the undamaged steel fracture toughness. and in most cases they were even greater than the undamaged steel fracture toughness (129. The overall average fracture toughness of A588-Beam-1 and A588-Beam-2 were equal to 258% and 204% of the undamaged steel toughness. and they reduced further for CVN specimens closer to the flange-web interface. and CVN specimens close to the flange-web interface had very low toughness values (approximately 12-22% of the undamaged toughness). which was subjected to smaller damage strains. The yield and ultimate stresses were slightly higher for the coupons (Z) closer to the flange-web interface. Only a few (3 out of 12) values were less than the AASHTO limit of 20 J. These values are within ± 10% of the corresponding undamaged A7 steel material properties. The change in ductility was similar for both A588 steel specimens. Each table includes the fracture toughness test results for the twelve CVN specimens (L1-1 to L1-4. The increase in yield stress was greater for specimen A7-Beam-1. Most of the fracture toughness values (11 of 12) were less than 25% of the undamaged steel toughness. They decreased consistently with distance from the flange edge. The overall average fracture toughness of A7-Beam-1 was about 20% of the undamaged steel fracture toughness. These values are within ± 10% of the corresponding undamaged A588 steel material properties for both beam specimens. and surface hardness of the A7 steel specimens. The damage-repair cycles reduce the ductility of the A7 steel specimens to about 80-96% of the undamaged steel ductility. These findings are consistent with those from phase-1 of the study (13. ultimate stress. This reduction is slightly greater for specimen A7-Beam-2. elastic modulus. which was subjected to lower damage strains. had much lower fracture toughness values than specimen A7-Beam-2. respectively. 14). Additionally. Several fracture toughness values were less than the AASHTO recommended limit of 20 J. and surface hardness of A588 steel specimens. The fracture toughness values less than the AASHTO recommended limit of 20 J at 4. and R1-1 to R1-4) from both specimens of the same steel type. and A36 steel beam specimens are presented in Tables 4. The damage-repair cycles do not have a significant influence on the elastic modulus. The toughness values were lowest at the center (C) Vee heat. Evaluation of Material Test Results for A588 Steel The normalized material properties in Table 3 indicate that damage-repair cycles do not have a significant influence on the yield stress. The toughness values were always greater than the AASHTO requirement of 20 J. Specimen A7- Beam-1. The fracture toughness values for specimen A7-Beam-2 were reasonable.75 J). A588. C-1 to C-4. For both beam specimens.

These findings are consistent with those from phase-1 of the study (13. The increase in yield stress and reduction in ductility were more significant for specimen A36-Beam-2. CONCLUSIONS The results from the experimental investigations and the material tests indicate that multiple (three) damage-heat straightening repair cycles do not have a significant influence on the elastic modulus (E). These values are within ± 10% of the corresponding undamaged A36 steel material properties for both beam specimens.e. and they decreased consistently with the distance from the flange edge. These results indicate that Tmax has a significant influence on the fracture toughness of A36 steel beam specimens. 14). After three damage-repair cycles. had slightly lower fracture toughness values than specimen A588-Beam-1. These findings are consistent with those from phase-1 of the study (13. but almost always greater than the fracture toughness of the undamaged A36 steel. These findings are also consistent with those from phase-1 (13. The average fracture toughness of A36- Beam-1 was about 84%. The three damage-repair cycles increase the yield stress and reduce the ductility of the A36 steel beam specimens. The three damage-repair cycles reduce the fracture toughness of the A7 steel beam specimen subjected to smaller damage strains significantly. Most of the fracture toughness values (with one exception) were greater than the undamaged steel fracture toughness (65. but some values were less than or close to the AASHTO limit. or A36 steels. The ductility of specimens A36-Beam-1 and A36- Beam-2 reduced to about 88-77% and 79-89% of the undamaged steel ductility. ultimate stress. The increase in yield stress is greater for specimens subjected to smaller damage strains. 9 . 14). The reduction in ductility is reasonable but not significant. respectively. A588. The results in Table 6 indicate that the fracture toughness values for A36-Beam-1 were highly variable. The toughness values were typically higher for the CVN specimens closer to the flange edge. and in some cases less than the AASHTO recommended limit of 20J. The average fracture toughness for specimen A36-Beam-2 was about 227% of the undamaged steel fracture toughness. Most of the toughness values were greater than the AASHTO limit of 20 J. and the lowest toughness values were about 22-33% of the undamaged steel toughness. Specimen A36-Beam-2 had much higher fracture toughness values than specimen A36-Beam-1. These findings are also consistent with those from phase-1 of the study (13. and they decreased consistently with the distance from the flange edge. The reduction in ductility is reasonable but not significant. i. 14). or surface hardness (HRB) of A7. 14). These results indicate that specimen A588-Beam-2. respectively.35 J). ultimate stress ("u). which was repaired with higher Tmax. The fracture toughness values were higher for CVN specimens closer to the flange edge. The damage repair cycles do not have a significant influence on the elastic modulus. The A588 beam specimen repaired with the larger Mr had slightly lower fracture toughness than the specimens repaired with smaller Mr.. The fracture toughness values for the A36 beam specimen repaired with higher Tmax (760°C) are also variable. The fracture toughness values for A36-Beam-2 were much greater than the AASHTO recommended limit of 20 J. The restraining moment is a relevant but not significant parameter for A588 steels. with several values less than the AASHTO recommended limit of 20J at 4. After three damage-repair cycles. The A7 steel beam specimen subjected to larger damage strains had reasonable fracture toughness values. The yield stresses of specimens A36-Beam-1 and A36-Beam-2 increased to about 107% and 110-121% of the undamaged steel yield stress. Evaluation of Material Test Results for A36 Steel The results in Table 3 indicate that the damage-repair cycles increase the yield stresses and reduce the ductility of the A36 beam specimens. The three damage-repair cycles do not have a significant influence on the yield stress or ductility of the A588 steel beam specimens. The three damage-repair cycles increase the yield stress and reduce the ductility of A7 steel beam specimens.4°C (40°F). the fracture toughness of the A588 steel beam specimens was much greater than the toughness of the undamaged A588 steel. the values after three damage-repair cycles are within ±10% of the undamaged values. and surface hardness of the A36 beam specimens. The increase in yield stress and reduction in ductility are more significant for the A36 beam specimen repaired with the higher Tmax (760°C or 1400°F). but they were also quite variable. 14). These findings are consistent with those from phase-1 (13. which was repaired using larger restraining moments Mr. the fracture toughness values for the A36 beam specimen repaired with lower Tmax (650°C or 1200°F) are quite variable.

. 2247-2262. H. 6. (1973). Avent. 9. PA.R. R.” Journal of Structural Engineering.J. Varma. 12. Heat-Straightening Repairs of Damaged Steel Bridges . Rothman. R.” E8. Report No. (2000).J.J. 15. PA. “Heat Straightening Prototype Damaged Bridge Girders. Vol. Kowalkowski. PA..” Journal of Structural Engineering. 12(3). 188-195. Ship Structures Committee. 235. 27-49. and Fadous. American Society for Testing and Materials. Dissertation. [computer file]. 14. and Robinson. AISC. 19. 283-296. 115(7). 13. 38(1). A. Federal Highway Administration.” E23." Engineering Journal. A. (1989). Report No. Vol. AASHTO (2004). Additional research is recommended to include the effects of realistic impact damage. 4. which are difficult to quantify in the field and implement in the laboratory. 5. D. of the Transportation Research Board Meetings. Guidelines for Evaluation and Repair of Damaged Steel Bridge Members.J. Effects of Multiple Damage-Heat Straightening Repair on the Structural Properties of Bridge Steels. “Multiple Heat Straightening Repair of Damaged Steel Bridges. “Standard Test Methods for Tension Testing of Metallic Materials. (1972). 11. local denting and buckling.F. Nichols.D. K.. 13(1). (1986).. D. 1631-1649. “Experimental Study of Heat Induced Deformation. Washington. (2001).R. W. Washington. The damage implemented in this research focused on weak-axis bending.. 137-141. 271. R. ASTM (2005).F. “Heat Straightening Damaged Steel Plate Elements. and Horn. Roeder. P. ASCE. 3. "What you Should Know about Heat Straightening Repair of Damaged Steel. R. P. and Robinson. 18-25. Lansing. D. and to further investigate the effects of overheating above 650°C.R. “Heat Straightening Rolled Shapes. FHWA-IF-99-004.” Engineering Journal.R. Oct. Shanafelt. pp. (2004). The research findings must be considered carefully before major decisions or recommendations. C. K. D. Michigan Department of Transportation. (2004).J. “Structural Properties of Steels Subjected to Multiple Cycles of Damage Followed by Heating Repair. Transportation Research Board. Avent. National Research Council. P.. Department of Transportation.G. IN. R.” Journal of Materials in Civil Engineering. Kowalkowski. LRFD Bridge Design Specifications – 3rd edition.” Ph.. and Robinson. and Kowalkowski.. RC-1456. Purdue University.J. Association of State Highway Transportation Officials. U.J. 18. REFERENCES 1.J. Avent. I. D. D. ASTM (2005). K.” Journal of Structural Engineering. R. L. West Lafayette.H..M.LIMITATIONS The major limitation of this study is that the testing approach does not account for the very localized effects of impact damage.” Journal of Materials in Civil Engineering.. 7. J. “Investigation of Triangular Heats Applied to Mild Steel Plates. ASTM (2005). American Society for Testing and Materials. 126(7) 755-763. P. and its similarity to the out-of-plane deformations of composite beam flanges in real bridge structures. 10. Mukai. The data compilation and report preparation were supported by Purdue University. “Effects of Multiple Damage-Heat Straightening Repair Cycles on the Structural Properties and Serviceability of Steel Beam Bridges. 16. R. 2. Report No.O.F. ACKNOWLEDGMENTS This research was sponsored by the Michigan Department of Transportation (Roger Till–cognizant official) and conducted using the laboratory facilities of Michigan State University. 133. (2005). and Varma.R. Avent. D. and Mukai. Avent. 2. ” Journal of Structural Engineering. (1998). G. Avent. “Residual-Stresses in Heat Straightened Steel Members. TRB.” Journal of Structural Engineering. and Boudreaux. (1984).J.. “Effect of Heat Straightening on Material Properties of Steel.F. . 747-754. No. 10 .W.. Varma A. Robinson. (2000)... West Conshohocken. R. American Socitey for Testing and Materials.H. “Standard Test Methods for Rockwell Hardness and Rockwell Superficial Hardness of Metallic Materials. R.J.” E18. Avent. NAS. (2007). R. Mukai. 17.C. 112(10). 126(7). Kowalkowski. and Shingledecker.” Proc. “Standard Test Methods for Notched Bar Impact Testing of Metallic Materials. (2000). 8.. Mukai. West Conshohocken. (2001). Flame Straightening Quenched and Tempered Steels in Ship Construction. K. G. pp. West Conshohocken. and Mukai. and Weerth.A Technical Guide and Manual of Practice. School of Civil Engineering. NCHRP Report No. Mukai. D. J. E.R.J.S.

and (b) after first damage cycle FIGURE 4 Schematic drawing of loading frame: (a) elevation. and (b) side view FIGURE 5 Plastic strain distribution in A7-Beam-2 subjected to damage strain of 90!y FIGURE 6 Vee heat locations and nomenclature (front side) FIGURE 7 Midspan displacements during all three damage cycles of A36-Beam 1 FIGURE 8 Experimental behavior during the second repair cycle of A7-Beam 1 FIGURE 9 Original locations of material samples in the flanges of damaged-repaired beam specimens 11 .LIST OF TABLES TABLE 1 Large-scale beam test matrix TABLE 2 Undamaged steel material properties TABLE 3 Normalized structural properties of damaged-repaired beam specimens TABLE 4 Fracture toughness of damaged-repaired A7 beam specimens TABLE 5 Fracture toughness results of damaged-repaired A588 beam specimens TABLE 6 Fracture toughness results of damaged-repaired A36 beam specimens LIST OF FIGURES FIGURE 1 Bottom flange displacements of steel beams impacted by over-height trucks FIGURE 2 Heat-straightening repair procedure used by MDOT Statewide Bridge Crew (SBC) FIGURE 3 Test setup and instrumentation layout for Specimen A7-Beam 2: (a) before.

25 125 53 53 650 A588-Beam-2 40 20 20 0.25 79 79 79 650 A36-Beam-2 30 30 30 0.25 79 79 79 760 12 .50 0.50 0. Cycle No.50 0.25 56 56 56 650 A7-Beam-2 90 60 60 0.50 0.25 0.50 125 53 53 650 A36-Beam-1 30 30 30 0.50 216 149 149 650 A588-Beam-1 40 20 20 0. Cycle No.50 0.25 0. 2. 3 A7-Beam-1 30 30 30 0.50 0. 1 2 3 1 2 3 1 2 3 1. TABLE 1 Large-scale beam test matrix !d / !y Mr / Mp-y # p (mm) Tmax (°C) Specimen ID Cycle No.25 0.25 0.50 0. Cycle No.25 0.

4 421.47 13 .35 78.64 106.3 210.32 A36 322.0 129.0 585.8 36.0 35.85 80.82 64.1 207.7 65.66 87.16 121.6 517.75 138.1 94.02 83.23 52.4 209. TABLE 2 Undamaged steel material properties Steel/Plate " yo Eo " uo eo HRBo Fracture Toughness (J) (MPa) (GPa) (MPa) (%) (B-Scale) Mean 95% H 95% L A7 252.34 A588 405.6 41.

16 0.07 0.94 1.01 0.04 1.00 1.98 0.15 1.89 A36-Beam-2 Y 1.97 0.98 0.69 0.99 0.88 X 1.98 0.94 1.94 Z 1.80 X 0.03 0. TABLE 3 Normalized structural properties of damaged-repaired beam specimens Specimen Coupon " y / " yo E / Eo " u / " uo e / eo HRB / HRBo X 1.96 0.92 0.03 0.94 1.20 0.00 0.02 0.99 0.15 0.07 1.95 Z 1.97 0.84 0.79 X 0.94 1.96 1.21 1.93 0.10 0.93 1.00 0.98 0.90 0.02 A588-Beam-2 Y 0.98 0.96 0.90 0.03 A588-Beam-1 Y 0.90 1.90 1.95 1.93 0.96 A7-Beam-1 Y 1.05 0.81 X 1.02 0.80 A7-Beam-2 Y 1.98 0.99 0.02 0.07 1.01 0.07 1.94 1.06 0.10 Z 1.25 0.04 0.08 0.81 X 1.07 Z 1.94 1.12 Z 1.97 A36-Beam-1 Y 1.79 14 .07 0.97 Z 1.96 1.16 0.

divided by 94. TABLE 4 Fracture toughness of damaged-repaired A7 beam specimens A7-Beam-1 A7-Beam-2 L1 C R1 Avg-H L1 C R1 Avg.07 0.19 0.37 2.13 0.84 0.85 1.88 0.03 1.20 1.14 0.-H 1 0..38 3 0.20 1.09 0.18 0.64 J) L1 C R1 Avg-H L1 C R1 Avg.04 0.e.73 0.56 1.12 0.12 0.68 4 0.11 0.23 0.40 0.05 15 .13 0.25 0.07 2.16 0.98 2 0.13 0.15 1.89 1.22 0.12 0.14 0.-H 1 69 12 24 35 191 100 268 187 2 22 7 14 14 228 79 84 130 3 18 14 12 15 98 11 83 64 4 11 9 11 11 12 20 17 16 Avg-V 30 11 15 19 133 53 113 99 Normalized fracture toughness (i.11 0.18 Avg-V 0.32 0.15 2.42 0.

06 3.e.08 3.72 2.36 1.05 2 3.36 1.08 3.09 0.24 4 1.98 Avg-V 2.08 3.26 1.06 2.75 J) L1 C R1 Avg-H L1 C R1 Avg-H 1 3.88 3 2.77 2.77 1.01 3.58 2. TABLE 5 Fracture toughness results of damaged-repaired A588 beam specimens A588-Beam-1 A588-Beam-2 L1 C R1 Avg-H L1 C R1 Avg-H 1 400 400 400 400 400 397 390 396 2 399 396 399 397 397 373 353 374 3 359 342 396 366 176 163 144 161 4 191 164 176 178 141 100 138 126 Avg-V 338 325 343 335 279 259 256 264 Normalized fracture toughness (i.37 1.07 3.07 0.08 3.60 2.25 1.04 16 .47 1.82 1.64 2.87 2.07 3.05 3.99 1.05 2.97 2.08 3.63 3.07 3. divided by 129..15 1.51 2.11 1.

94 Avg-V 0.37 1.35 J) L1 C R1 Avg-H L1 C R1 Avg-H 1 1..18 2.77 2.35 0.25 1.77 1.28 2.95 3.51 0.73 3.30 2 1.33 1.32 2.84 1. TABLE 6 Fracture toughness of damaged-repaired A36 beam specimens A36-Beam 1 A36-Beam 2 L1 C R1 Avg-H L1 C R1 Avg-H 1 92 22 132 81 125 278 244 216 2 89 92 50 77 152 193 255 199 3 23 22 87 43 125 75 146 115 4 14 15 22 18 61 24 99 61 Avg-V 54 38 73 56 115 142 186 148 Normalized fracture toughness (i.91 1.01 1.26 0.67 1.90 3.e.25 3.37 1.84 0.41 0.41 0.33 0.23 0.27 17 .11 0.33 0.33 2.14 2.24 1. divided by 65.85 2.93 0.06 3 0.76 4 0.58 1.22 0.91 4.

FIGURE 1 Bottom flange displacements of steel beams impacted by over-height trucks 18 .

FIGURE 2 Heat-straightening repair procedure used by MDOT Statewide Bridge Crew (SBC) 19 .

and (b) after first damage cycle 20 .FIGURE 3 Test setup and instrumentation layout for Specimen A7-Beam 2: (a) before damage.

and (b) side view 21 .FIGURE 4 Schematic drawing of loading frame: (a) elevation.

FIGURE 5 Plastic strain distribution in A7-Beam-2 subjected to damage strain of 90! y 22 .

FIGURE 6 Vee heat locations and nomenclature (front side) 23 .

FIGURE 7 Midspan displacements during all three damage cycles of A36-Beam 1 24 .

FIGURE 8 Experimental behavior during the second repair cycle of A7-Beam 1 25 .

FIGURE 9 Original locations of material samples in the flanges of damaged-repaired beam specimens 26 .