Shape control by local heating is often used in the processing of structural steels. Experience has proven that normalized steels are suitable for this operation if it is carried out properly. In recent years, thermomechanically controlled rolled (TM) steels have been developed. For their application, it was necessary to investigate the response of this steel type to flame straightening. Two essential processing conditions must be considered: heating limited to the surface and heating of the full material thickness. The temperatures were varied from 550°C to 950°C through different heat inputs. Tensile and Charpy impact tests have proven that within this temperature range the investigated steels (S355ML-EN10113) were not affected. For conditions of wedge heating, a decline in mechanical properties was observed if the temperatures exceeded 650°C. Both limits 950°C (in line heating) and 650°C (in wedge heating) allow for the effective shaping of steel constructions. Experienced operators can easily respect these limitations. Hence, the TM rolled heavy plates are suitable for flame straightening.
In the fabrication of steel structures, flame straightening is often used in order to adjust the geometry to the dimensional tolerances. Flame straightening (or forming) has proven its value in many areas of steel processing. Skilled operators can reach extremely close tolerances. The process can be used without heavy mechanical equipment, which is an important advantage, especially for onsite operation. Many practical hints are given by Pfeiffer.
Normalized structural steels have not shown particular problems after flame straightening, and the process has been generally accepted for structural steel work, ship building and other applications. In the past few years, TM steels have become increasingly used for the same purposes thanks to their improved weldability. However, softening was expected when heating TM steels to temperatures exceeding 580°C. Hence, this temperature was set for all fabrication processes, including flame straightening.
Low temperatures hardly allow for an effective shaping. Therefore, it was of industrial importance (in regards to the use of TM steels) that higher temperatures are acceptable as well. The investigation included herein was carried out in order to prove this.
Flame Straightening Experimental Procedure
The investigation was carried out on plates of 355 MPa nominal yield strength (corresponding to API 2W Grd 50and ENI 0113 S355ML). Three plates (15, 50 and 75 mm thick) were taken from the plate production of Dillinger Hütte GTS. The parent plate tensile properties given in Table 1 confirm that with TM rolling very lean chemical compositions (Table 2) are sufficient to meet the requirements even for important plate thickness. The 15 mm plate represents a typical S355ML, rolled-for-structure steelwork. For such plates a simple TM rolling is sufficient to achieve the requirements. For thicker plates, accelerated cooling after the final rolling pass is beneficial to ensure an optimized grain refinement. Plates of relatively high tensile properties were selected for the test program as it was expected that such plates would react most sensitively to softening.
Line Heating of Steel Plate Test Panels
Through line heating of thick plates the thermal field is mostly concentrated close to the plate surface, as schematically shown in Figure 1.
As direct measurements of the temperature at the surface exposed to the flame are not possible, an indirect measurement technique had to be adopted. The test arrangement is shown in Figure 2.
Temperatures were measured through the use of thermocouples that were introduced from the bottom side and fixed by TIG-welding at about 2 mm subsurface. It became obvious from the microstructure that the temperature was 50-150°C higher at the plate surface. However, in the paper the measured values are referenced as flame straightening temperatures Tmax.
Line heating was applied to a number of test panels cut from 15 and 50 mm thick plates. A multiflame oxyacetylene torch was used. At first a correlation between travel speed and maximum temperature (Tmax) was established for constant gas flows. The travel of the torch was mechanized to ensure constant heat inputs. The traveling speed of the torch was then adapted to produce maximum temperatures of 650, 750, 850 and 950°C.
For a special set of tests the flame straightening was repeated, affecting the material three times with the same Tmax. For the standard single cycles and the triple cycles the heated areas cooled through free heat flow. In practice, water cooling is sometimes also applied to force the straightening effect and to accelerate the operation. Such conditions were therefore included in the test program as well. The water shower followed the torch at such a distance that the temperature at the point of water cooling was below 600°C.
Tensile and impact specimens were machined from the panels and polished and etched surfaces prepared for metallographic examination. Full thickness flat tensile specimens were tested to look for the integral behavior of the shaped plates. More localized effects of flame straightening on the mechanical properties were characterized in testing subsized specimens extracted at the surfaces. These flat tensile specimens were 3 mm thick for the 15 mm plate and 5 mm for the 50 mm plate. The parent plate properties were tested parallel on specimens of the same type to allow for a direct comparison. For impact tests, standard Charpy specimens were used located subsurface.
Thermal Simulation Tests
Specimens subjected to time-temperature cycles corresponding to flame straightening were tested in order to measure more local properties. The heat treatment was performed using conductive heating in a Gleeble equipment. The full cross-section of the simulated specimens consisted homogeneously of a microstructure corresponding to a particular area of the heat-affected zone.
The following thermal cycles were applied: heating in 60 s to maximum temperature, holding time 60 s, cooling according to t8/5 cooling time of 20 s. Peak temperatures of 650, 750, S50 and 950°C were chosen for the thermal cycles.
Single cycle and triple cycles repeating the same thermal conditions were used. From the simulated bars Charpy impact and small size tensile specimens were prepared.
The third part of the investigation dealt with the conditions of shape control, where approximately the same temperature is reached through the full plate thickness. Such conditions are typical for wedge heating to straighten girders or point heating for the flattening of buckles. These procedures generally use softer flames or waving. The heat spreads across a larger area, heating and cooling are much slower compared to line heating and the steel is exposed longer to the peak temperature.
Full thickness heating was simulated by simple furnace heat treatments. Panels of 15, 50 and 75 mm thickness were exposed to different temperatures. The holding times were adapted to the thickness of the plates. A two-minute holding time per mm plate thickness should represent the worst-case scenario during straightening operations. After removal from the furnace, the panels cooled in still air.
Flame Straightening Test Results
Effect of Flame Straightening on Tensile Properties
The influence of different flame straightening conditions on tensile properties are shown in Figure 3 for the 15 mm plate, and in Figure 4 for 50 mm panels. Each symbol in the figures represents one individual test result, so an inherent scatter has to be accepted. No clear trend can be seen as a function of maximum temperature from 650°C to 950°C, and the level remained unchanged in regards to the parent plate properties. The slight differences can rather be attributed to inherent scatter rather than to metallurgical effects.
It must be noted that the depth of the heat-affected zones was small in relation to the specimen thickness. Thus, a strong softening would have been necessary to show a remarkable global softening in testing the full thickness specimens.
In order to get information about more local properties, flat tensile specimens of reduced thickness were tested. This part of the investigation may be more relevant for metallurgists than for designers or users of the steels. In regards to the parent plate properties, the determined subsurfaces were identical to the full thickness specimens. Yield strength between 420 and 450 MPa were measured for the 15 mm plate, and 390 to 420 MPa for the 50 mm plate. This shows that the TM rolling created a homogeneous fine-grained microstructure across the plate thickness and no surface hardening was produced by the accelerated cooling of the 50 mm plate.
After flame straightening, the subsurface yield strength was slightly raised, reaching an average of 450 MPa for all test conditions. The ultimate tensile strength was also increased by some 40 MPa compared to the values determined on the full thickness specimens.
Effect of Flame Straightening on Impact Toughness
The toughness within the different zones was characterized by Charily impact tests. The specimens were machined from the subsurface and orientated transverse to the plate rolling direction. They were fractured at different temperatures so that full transition curves could be established. The curves of the 15 mm panels in the parent condition and after flame straightening to different maximum temperatures are shown in Figure 5. The lines within the graph connect the average values, each representing a set of three individual tests per series. No systematic influence of the flame straightening on the impact toughness could be observed.
Further results from triple cycle and water cooling are summarized in Table 3 along with the corresponding results of the 50 mm panels. In order to characterize the transition behavior, the temperatures (with an impact energy of 50 or 100 J) were determined. This was done by using an eye fit average curve. The transition temperatures (TT50J and TT100J) confirm that excellent impact toughness was obtained for all test conditions. This is not only true for the single cycles, but also includes multiple treatments and accelerated cooling by water shower. A slight shift of the brittle fracture transition curve may be interpreted for some of the curves for specimens that have been subjected to inter-critical heating (max. temperatures 750-850°C). The difference was below 15°C, if any at all. The most brittle impact series reached 50 J at temperatures as low as -79°C.
One can therefore conclude that for the thermal conditions of line heating, low carbon TM-steels were not susceptible to embrittlement.
Impact Test on Simulated Microstructures
So far only impact results were reported, whereas the specimen HAZ areas heated to different temperatures were sampled so that an integral fracture behavior of different microstructures was determined. Information about the toughness of individual microstructures of the heat-affected zone was assessed on thermally simulated specimens.
The impact transition temperatures, hardness results and tensile properties from this investigation are summarized in Table 4. The following trends were observed:
The hardness was not significantly changed for any of the test conditions. The tensile strength was also not altered, and the results were within a close range (between 533 and 550 Mpa). For a 0.2% yield strength, a minimum was obtained. Due to partial transformation, carbon was locally enriched during the intercritical heating with a peak temperature of 850°C, and some ferrite grains were observed for which the grain size (ASTM) had increased from 10-11 (parent plate) to 9-10. Compared to single cycles the specimens subjected to triple cycles led to about 30 MPa lower yield strength for all cycles (again, as a result of more pronounced carbon segregation).
Impact transition temperatures were at an excellent level for any of the thermal cycles—with the worst TT50J at -76°C (triple cycle)—and exceeded by far the requirements for the parent material. As for the yield strength, the slightly reduced toughness can be explained by increased ferrite grain size and carbon concentration in the areas of the microstructure that were transformed to austenite at peak temperature.
Tensile Properties/Mechanical Properties for the Conditions of Wedge Heating
For temperatures commonly used for PWHT and up to 625°C, no change of properties was noted for any of the tested plates. In order to separate more distinctively the effects caused by tempering from those caused by a partial transformation, temperatures of 700 and 720°C were also tested, which is just below the transformation temperature. It was observed that this treatment effected a globularization of cementite particles so that the yield strength was reduced by approximately 30 MPa.
Figure 6 shows that the yield strength was again lowered when higher temperatures were applied. Slower cooling and hence slower transformation led to a remarkable increase in ferrite grain size compared to the parent material. As the cooling rate for air cooling decreased with increasing plate thickness, the 75 mm plate had the lowest cooling rate, the biggest grains and consequently the lowest yield strength. For conditions of partial austenization a softening was obtained, which could be attributed to the formation of selective larger ferrite grains and carbon segregation. The drop in Y.S. compared to the parent plate reached about 50 MPa, so that after 750°C heat treatment the requirements were hardly fulfilled. Higher temperatures leading to complete transformation again led to somewhat higher properties, but still below the initial values. Compared to the behavior after Tmax in the intercritical range, carbon distribution was more homogeneous and absence of individual ferrite grains of excessive size was obtained.
Discussion of Flame Straightening Test Results
Due to the high cooling rates, a direct control of actual maximum temperature at the surface is impossible. The procedure therefore requires experienced operators who are able to control temperature by distinguishing the color of the heated spot.
Thanks to the lean chemical composition, hardenability of the TM steels is low. A quickly cooled HAZ consisted of fine-grained mainly ferritic microstructure. The tensile properties were not significantly changed despite a wide range of applied maximum temperatures. No softening occurred under any of the test conditions. For some test conditions a local increase in yield and tensile strength was obtained but the effect was limited to the very surface and only reached some 10%.
As long as no temperature was reached that caused excessive austenite grain growth, no local embrittlement was observed. It is known from welding that large austenite grains are generally limited to areas that have been heated to more than 1000°C. Consequently, 950°C should be acceptable and still provide a certain safety margin.
When softer flames are used or larger areas heated (e.g., by waving the torch), it takes longer to reach the required surface temperature. Whilst heating, the heat has time to spread deeper into the surrounding material and the thermal gradient becomes less pronounced. This leads to retarded cooling when the flame is taken away. At the same time a more important part of the cross-section may be affected so that more caution is needed to guarantee the full load bearing capacity of the construction. High speed pyrometers or appropriate contact thermocouples can be used to control the temperatures during the operation. The operator should calibrate the procedure on test plates in order to be able to avoid excessive temperatures. Hardness tests on the treated surface could be used for quality control in the final condition.
Slowest and longest heating cycles are obtained when the full plate or flange thickness is homogeneously heated to temperature. This situation, which is typical for wedge heating, was simulated by furnace heat treatment.
The lowest tensile properties were obtained after a selective austenizing at 750°C followed by air cooling. The yield strength was lowered by some 50 MPa. It is worth noting that the tensile strength was hardly altered. The yield strength suffered from single large ferrite grains caused by the slow selective transformation. Compared to the applied furnace heat treatment, wedge heating in practice generally will have faster heating due to the direct impingement of the plate surface by the flame. Furthermore, the time of exposure at high temperatures will normally be much shorter. Both differences should produce fewer critical material properties compared to the tests. In considering these critical conditions of the test conditions, the softening of 50 MPa is relatively small. The parent plate requirements could still be met for the tested plates. Consequently, even after these critical conditions it can be expected that, after a certain hardening, the TM steels would stand full design stress.
To avoid any risk, the temperature for wedge heating should be restricted to temperatures clearly below the transformation temperature. Higher temperatures would at the same time reduce the straightening effect.
Conclusion: Thermomechanically Rolled Heavy Plates can be Flame Straightened
The tests carried out on steel grade S355ML have shown that thermomechanically rolled heavy plates can be flame straightened without particular difficulties.
In line heating, a surface temperature of at least 900°C can be tolerated without detrimental influence to mechanical properties. An experienced operator can follow this limit by controlling the color of the hot spot directly after removing the flame.
In case of a flame straightening procedure where the full thickness of the plate is heated and hence slower cooling is obtained, temperatures should be restricted to 700°C to avoid any risk of local softening. If longer holding times apply, the temperature limit should be reduced to 600-650°C. Since heating and cooling happen much slower for this technique, the temperature can be reliably controlled with contact thermocouples, pyrometers or temperature-indicating crayons.
The results of the investigation presented in this paper allow us to assume that within the above process limits the materials properties are not affected.
TM steels of very high yield strength (e.g., S690M) may be more prone to softening or embrittlement by flame straightening. They may therefore require closer temperature ranges.
This article originally appeared at Dillinger Hütte and is reproduced with permission.