Mitigation of Weldment Cracking in Steel Highway Structures Due to the Galvanizing Process (2021)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

65   Summary of Research Findings This chapter summarizes the major findings from this research. These findings were derived for the material grades and bath compositions used in the research; different materials, bath chemistries, or environments may lead to different findings. In this research, LMAC—as measured by the specimen’s strain at failure—was used as an indicator of susceptibility of steel specimens to cracking during galvanizing. Therefore, the research findings and conclu- sions were based primarily on strain-at-failure data. Experimental Investigations Steel Material Specimens of all grades of steel included in the study (A36, A572-50, A572-65, and HPS100W) experienced larger reductions in total elongation when tested in the galvanizing bath than when tested in air at RT or ET. The following are specific findings: 1. All steels included in the test program experienced reductions of strain at failure in the presence of zinc. The reductions in strain for the in-zinc specimens compared to the in-air at RT specimens averaged 65% for A36, 57% for A572-50, 58% for A572-65, and 70% for HPS100W (when no other factors were included). 2. A36, A572-50, and A572-65 steel specimens exhibited similar magnitudes of strain at failure, averaging about 8% when tested in zinc; HPS100W specimens exhibited much lower strain at failure, averaging about 4% when tested in zinc. 3. Research findings indicate that the galvanizing process contributes in varying degrees to the reduction in strain values at failure for the different steel grades because the strains for steel components in the bath accumulate from various sources. The likelihood of expe- riencing LMAC will decrease as the strains at failure in the bath increase. A36, A572-50, and A572-65 steels exhibited similar average strains at failure that were higher than those exhibited by HPS100W steel. It should be recognized that there are considerable nuances in these findings to differen- tiate between A36, A572-50, and A572-65. Also, other grades of steel are used in highway structures—such as A500 and A595, which are commonly used for tubular structural elements— as well as HPS70W and HPS50W; these were not considered in this research. Galvanizing Bath Chemistry Four different bath chemistries commonly used in the U.S. galvanizing industry were consid- ered in the test program: SHG, SHG + 0.1% Bi, SHG + 1.0% Pb, and SHG + 0.1% Bi + 1.0% Pb. Tension tests were conducted in each of the bath compositions for the four material grades; tests C H A P T E R 6 Summary and Recommendations for Research

66 Mitigation of Weldment Cracking in Steel Highway Structures Due to the Galvanizing Process were conducted in SHG and SHG + 0.1% Bi + 1.0% Pb for notched specimens, HAZ-simulated specimens, and cold-worked specimens. Test results yielded the following findings: 1. Galvanizing bath composition did not affect average strains at failure for all material grades when no other factors were present. 2. When a stress concentration (notch) was introduced, the presence of bismuth and lead in the galvanizing bath resulted in smaller reductions in strains for A36 and A572-50 than those for tests performed in SHG. The notched HPS100W specimens resulted in such small elonga- tions in zinc that no specific conclusions regarding notch severity are drawn; no similar tests were performed on A572-65 specimens. 3. When bath chemistry and steel grade were considered in combination with HAZ properties and a notch (all HAZ-simulated specimens were notched), lower strains were observed for A36, A572-50, and HPS100W when bismuth and lead were included in the bath; these elongations were lower than those resulting from just a notch. 4. When the bath chemistry and steel grade were considered in combination with cold-working, no observable difference in strain was observed for CW6 specimens when bismuth and lead were added to the bath. However, a slight decrease in strain was consistently observed for the CW8 specimens of all steel grades when bismuth and lead were added to the bath compared to those tested in pure zinc. 5. Test results indicated a tendency of Bi + Pb to aggravate LMAC in steel structures during galvanizing in the presence of stress concentrations, HAZ, and/or cold-working. While the differences in strain between specimens tested in SHG and those tested in SHG + 0.1% Bi + 1.0% Pb were not large, it is likely that the effect of bath chemistry will increase as the effects of these factors are combined, as observed when comparing the results for the notched, HAZ-simulated specimens with those for notched, non-HAZ specimens. Stress Concentration Two different levels of stress concentration were considered in the study, achieved by notch- ing tensile specimens with either SCF = 1.5 or SCF = 2.5. The following conclusions were based on a comparison of the strains at failure for notched specimens tested in zinc to those for notched specimens tested in air. 1. The specimens with more severe notches experienced larger percentage reduction in strain between the in-air and the in-zinc cases than specimens with less severe notches, indicating an effect of the severity of stress concentration on LMAC. The level of this effect was similar for A36 and A572-50 specimens (A572-65 steel was not included in these tests). However, this effect was not observed in HPS100W, and the strains for the notched specimens tested in zinc were too small to distinguish any difference. 2. Test results indicated a significant effect of severity of stress concentration on LMAC, sug- gesting that severe stress concentrations will increase the likelihood of LMAC occurrences in steel structures during galvanizing. The SCFs used were not exceptionally severe to ensure that the data would represent a practical situation (e.g., a hole in a plate presents an SCF of approximately 3.0). Therefore, such an effect would become more pronounced for weldments and connection details with severe SCFs and for structures that have sharp or crack-like flaws. HAZ Properties Three different HAZ microstructures, as indicated here by cooling rates, were considered: SCR = 12°C/s (corresponds to 60 kJ/in. of heat input), MCR = 30°C/sec (corresponds to 30 kJ/in. of heat input), and FCR = 60°C/sec (corresponds to oxyfuel flame cutting on a 2 in. plate). In addition, HAZ-simulated specimens were tested in air at RT after acid pickling to investigate the hydrogen embrittlement effects caused by these factors. The HAZ-simulated specimens

Summary and Recommendations for Research 67   were 252-type, not the 505-type used for investigating, and were notched to ensure failure at mid-length. Thus, the results reflect both the influences of notching and HAZ microstructures. The following are the findings of these tests: 1. All specimens experienced extremely low levels of strain in zinc. 2. The specimens subjected to an FCR were more susceptible to LMAC than those subjected to an MCR or SCR. 3. There was a clear effect of the galvanizing bath chemistry, showing significantly lower strains at failure when lead and bismuth were added to the bath. 4. HAZ has a higher effect on the reduction of strain at failure for HPS100W than for A572-50, and it has a higher effect on A572-50 than on A36. These effects were only noticeable when bismuth and lead were included in the galvanizing bath, and they were so severe that all HAZ-simulated HPS100W specimens tested in SHG + 0.1% Bi + 1.0% Pb failed at strains less than 0.2%. 5. The HAZ-simulated specimens tested in air after acid pickling performed similar to unpickled HAZ-simulated specimens tested in air, indicating that pickling-induced hydrogen embrittle ment is not an influencing factor. 6. All specimens subjected to an FCR exhibited the greatest hardness and showed the greatest reduction in strain at failure. There was also a general trend of sincreasing material hardness with increasing strength. 7. The combination of an FCR and stress concentration was found to produce high reductions in strain at failure; these factors represent the condition of oxyfuel flame-cut surfaces in practice. Also, FCRs resulted in lower strains at failure in A572-50 steel specimens than in A36 steel specimens. Cold-Working Three different levels of cold-working were considered: Two levels were obtained through direct tension pre-straining to 6% and 8%, while the third level was obtained by cold-bending A572-50 plate 90° and extracting tension specimens along the bend line. Thirty-nine specimens were pre-strained in tension to 6% or 8%, and the strains at failure in zinc were compared to those obtained for non-cold-worked specimens tested in zinc and for cold-worked specimens tested in air. The following observations were made: 1. The strains at failure of CW6 and CW8 specimens tested in zinc were comparable to those for non-cold-worked specimens tested in zinc, indicating that cold-working is not a major contributing factor to LMAC. 2. A slight effect of bath chemistry on the CW8 specimens was noted. The specimens exhibited small reductions in average strain at failure when lead and bismuth were added to the zinc compared to the strains for tests conducted in SHG. No similar effect was observed for the 6% cold-work level, indicating that increasing levels of cold-work may contribute to a slight increase in susceptibility to LMAC. 3. Cold-working did not greatly affect the material hardness of the specimens. In general, it appears that cold-working is less of a contributing factor to susceptibility to LMAC than the flaws introduced during the cold-bending process. Material Performance in the Presence of a Sharp Crack Fourteen WOL specimens—made from A36, A572-50, A572-65, and HPS100W steel—were tested in the galvanizing environment. Each specimen included a sharp crack formed under fatigue loading, and each was self-loaded when submerged in the bath such that the crack tip was subjected to a tensile stress. None of the specimens exhibited any crack growth in the liquid zinc environment, indicating that exposure to liquid zinc does not contribute to crack growth.

68 Mitigation of Weldment Cracking in Steel Highway Structures Due to the Galvanizing Process Computational Simulations Thermal-mechanical analysis of 19 HMIPs was performed to evaluate the effect of struc- tural configuration on nonlinear stresses and strains during the HDG process. The models were calibrated to thermal measurements taken from an HMIP structure during galvanizing, and a parametric study was performed to investigate the influence of pole thickness, base plate thick- ness, and pole shape (round versus polygonal; number of sides) on strain and stress responses. The computational simulations indicated the following: 1. Poles experienced the largest stresses/strains upon contact with liquid zinc during the dipping part of the cycle, and another peak upon extraction from the zinc and contacting air. 2. Pole cross-sections did not deform uniformly during dipping; a complex mode of deforma- tion behavior was observed. Different locations around the pole circumference were most susceptible to LMAC during the galvanizing process. 3. As the ratio of base plate thickness to pole thickness increased, stresses and inelastic strains increased at the base plate connection. This behavior occurred because the pole expands more rapidly than the thicker base plate upon contact with zinc, and the base plate restrains the expansion of the pole at the connection. In other words, as the thermal mismatch between the base plate and pole increases, strains and stresses at the connection detail becomes more severe. This finding indicates an increased susceptibility to LMAC as the ratio of base plate thickness to pole thickness increases. 4. Inelastic strain and stress concentrated significantly at bend lines in multisided poles. The level of concentration was greater for larger ratios of base plate thickness to pole thickness. 5. Multisided poles consistently experienced greater inelastic stresses and strains during the galvanizing process than round poles. The models did not capture the effects from material cold-working or residual stress, which would increase the stresses and strains more for mul- tisided poles than for round poles. Suggested Revisions to Current Specifications Based on the research findings, changes to AASHTO LRFD Specifications for Structural Supports for Highway Signs, Luminaires, and Traffic Signals, first edition (2015) and AASHTO LRFD Bridge Design Specifications, eighth edition (2017) were proposed; these are provided in the attachment. These changes are expected to improve the design, materials, and construc- tion specifications of galvanized steel highway structures and help mitigate weldment cracking caused by the galvanizing process. Recommendations for Future Research This research addressed the influence of a number of factors, individually or in combina- tion, on susceptibility of steel highway structures to LMAC. These factors were steel grade, galvanizing bath chemistry, stress concentration, cold-working, and HAZ properties; other factors can be evaluated using the same testing approach. Also, nonlinear computational simulations were executed to provide insights into the effects of geometric aspects of HMIP structures. This work has provided an understanding of those factors contributing to LMAC in steel structures during galvanizing. The following topics are proposed for research to further help control cracking associated with galvanizing of steel highway structures: 1. This research highlighted the interaction between stress concentration and galvanizing, suggesting the role of connection detail geometry in controlling LMAC during galvanizing. Future research that includes both full-scale experimental tests and computational simula- tions is needed to address susceptibility of specific connection details to LMAC.

Summary and Recommendations for Research 69   2. The research program showed observable and consistent interactions between galvanizing bath chemistry (SHG + Bi + Pb) and other factors such as HAZ properties, stress concentra- tion, and cold-working. It would be helpful to establish, through research, the interactive effects of other galvanizing baths, such as tin. 3. Three of the four steel grades studied in the project (A36, A572-50, and A572-65) performed similarly in zinc; the other grade (HPS100W) performed differently, showing more inter- active effects with the galvanizing bath chemistry and other factors. Further research on other grades of steel, such as HPS70W, will help establish the effects of the galvanizing pro- cess on the performance of other steel grades used in highway structures.