7. RESULTS AND DISCUSSION
7.5. Surface crack detection
All visual surface crack detection tests were made using penetrant testing in accordance with testing standards SFS-EN 571-1 and SFS-EN ISO 23277. Any crack detections were observed and the size of the undercut was within in the limits of the standard as the welding had been conducted in a laboratory environment.
104 7.6. Transverse tensile test
Two transverse tensile tests were performed on all welds in accordance with standard SFS-EN ISO 4136. The results of these tests are in figs 40 though 43.
Fig. 40 represents all of the tensile test results that were collected and helps to illustrate that the tensile strength of the welded structure is lower when the heat input is bigger. The tensile strength of the filler material was 560 MPa, and the tensile strength of the base material, corresponding to its material standard, was between 700 and 770 MPa. All of the material certificates have actual values of tensile strength. When undermatched filler metal was used during welding, the real tensile strength of the undermatched welded structure was more than the tensile strength of filler material as a consequence of penetration and mixing between the base and filler materials. The tensile strength of the welded structure is near the tensile strength of the base material required by that steel’s standard. All of this can be seen in figs 40 through 42 and additionally all of the tested welded structures broke at their welding points as a result of the tensile test.
Heat input 1.0 kJ/mm Heat input 1.3 kJ/mm Heat input 1.7 kJ/mm
Tensile test values of structure
A B C D E F G H
MPa
STEELS
Figure 40. Tensile strengths of welded joint made of different steels using three heat input.
105 Tensile strength of filler material 560 MPa TENSILE STRENGTH OF WELDED STRUCTURE
STEELS 780 MPa
Figure 41. Tensile strength of various steels using constant heat input 1.0 kJ/mm. Tensile strength of filler material 560 MPa TENSILE STRENGTH OF WELDED STRUCTURE
STEELS 780 MPa
Figure 42. Tensile strength of various steels using constant heat input 1.3 kJ/mm.
106 Tensile strength of filler material 560 MPa
STEELS
TENSILE STRENGTH OF WELDED STRUCTURE
780 MPa
Figure 43. Tensile strength of various steels using constant heat input 1.7 kJ/mm.
The mismatch level between filler metal and parent metal was 0.72, which is lower than the recommendations of many researchers (Toyota 1986, Satoh & et al. 1975). Within such as low mismatch value, it is clear that the weld is the weakest place in structure, especially when compared to the strengths of filler and base materials. Tensile strength values change when penetration and mixing between filler and base material occurs and figures 41 through 43 show that the strength values of the base material are higher than the filler material. A typical example of a broken tensile test bar is shown in figs 44 a, b, c and d.
The fracture occurs in the weld at the point of reduction area, the failure of which arises in the HAZ and continues into the weld.
107
a)side picture b) face side
c) root side d) broken tensile test bar Figure 44. Tensile test bar.
Steel A has standard tensile test value 700 MPa and using the lowest heat input (1.0 kJ/mm), the values obtained from steel A were near the tensile strength of the base steel. The same happened in steel H when heat input was 1.0 and 1.3 kJ/mm, and nearly same situation occurred in steel D. In these situations, the tensile strength of welded structure was 4 % lower than the tensile strength of base material. The standard tensile strength of all steels with the exception of steel A was 780 MPa.
In all welded structures, the failure started from the weld or the HAZ, however, in some instants the failure started from the fusion line between weld and HAZ.
This happens because of the low yield strength of filler material but also because of possible deformation in the weld. Nearly all of the tensile test pieces failed starting at the fusion line and only few of them broke in the HAZ. When the failure began in the fusion line or the HAZ, the direction of the break was towards the weld at a traditional 45° angle.
108
There will always be differences between welded structures regardless of how the steel was welded or what the heat input of welded structure was. All of heat inputs, steel H had the best tensile strength values, closely followed by steel G.
In all cases, tensile strength values were the lowest when the heat input was 1.7 kJ/mm, however steel B’s lowest tensile test occurred when the heat input was 1.0 kJ/mm. It is important to consider that tensile test values do not account for all feature of the welded structure, and this is why other tests were conducted within the scope of this research to determine other mechanical properties.
Tables 29, 30 and 31 present tensile test values which are used in different heat inputs in welding. Regardless of heat input, all tensile test values are higher than the tensile test of the filler material, which was 560 MPa. The tensile test values of the base material was around 780 MPa or more (steel A had minimum tensile test value 700 MPa). In all steels, the real tensile test value was more than in manufactorer’s procedure. Heat input has lowering effect to tensile strength of structure. Manufactory method doesn’t effect to tensile strength.
Also, TMCP and QT steels behaved equally when using different heat input in welding.
In all structures, the elongation at the break was considerably smaller than the base material elongation. In 690 MPa class HSSs, the standards stipulate that the minimum A5 should be 15%. However, in this research the values for the elongation at the break were only half of the base material values. These discrepancies can be accounted for by the differences in elongation at the break between the base and filler material as seen in tables 29, 30 and 31. The gauge length was 85 mm (standard SFS-EN ISO 6892-1) while the length of the weld was around 25 mm. The yield strength of the filler material was 470 MPa while the yield strength of the base material was 690 MPa. As there was such a large difference between these yield strength values, most of the yielding occurred in the weld. As these steels were constructed under varying manufacturing methods, their resulting yield strength and elongation break
109
values differed from one another. These differences caused variations between elongation break values of the welded structures when using the same heat input. The amount of penetration and dilution that occurred between the base and filler materials led to a better tensile strength in the welded structure than in the filler material. There is a correlation between the tensile strength and elongation break value of HSS, where larger real tensile strength leads to smaller elongation break values.
When dilution happens between the base material and the weld, alloy elements can mix together. Some alloys such as Nb mixes to the weld and increases the properties of the welded structure. The Metal Handbook (1990) explains that the yield strength of the carbon steel increases with small additions of Nb. The yield strength of carbon steel can increase from 490 MPa to 700 MPa when the addition of Nb is 0.02 %.
Using fillet welds, it is possible to increase the size of the weld (effective throat thickness) which leads to a greater tensile strength in the welded structure.
Aside from increasing the tensile strength, this method also has some negative side effects including a longer welding time, higher cost and decreasing productivity. As opposed to fillet welds, butt welds are limited and increasing the weld size is not possible. When using undermatched filler material, the welded structure will not have a strength matching its base material.
110
Table 29. Comparing the tensile strength and elongation at break of base ma-terial to the welded structure when heat input was 1.0 kJ/mm. Red font corres-ponds to the highest value while green font correscorres-ponds to the lowest.
Table 30. Comparing the tensile strength and elongation at break of base material to the welded structure when heat input was 1.3 kJ/mm. Red font corresponds to the highest value while green font corresponds to the lowest.
TEST
111
Table 31. Comparing the tensile strength and elongation at break of base ma-terial to the welded structure when heat input was 1.7 kJ/mm. Red font corres-ponds to the highest value while green font correscorres-ponds to the lowest.
This tensile test has proven that when heat input is bigger and consequence of that width of HAZ is consequently wider, the tensile properties of the welded structure are weaker than the base material. In the tensile tests, the weakest welded structure had the highest heat input. Rodriques et al. (2004a) came to the same conclusion in their study when they looked at matched and under-matched filler metal situations and determined that the strength of the joint is strongly depend on the HAZ dimension. It is therefore of utmost importance to use proper welding parameters when welding HSSs regardless of the filler ma-terial.
112 7.7. Transverse bend test
Overall, four bend tests were carried out to determine the occurrence of cracks and unmelted fusion line among other issues. Two of these tests were carried out on the root of the groove, while the other two were carried out on the top of the groove. All bend tests were made according to standard SFS-EN ISO 5173.
The transverse bend tests will show faults in welded structure, such as defective penetration or low mixture levels between base and filler material.
The transverse bend tests that were done on these HSSs with undermatched filler material were much more demanding than normal transverse bend tests.
The discrepancy between the tests occurs because the filler material has a lower yield strength than base material. In these tests, the first part to be bent was the welded structure and the base material. In the end of these tests, the weld yielded more than the base material and the bending angle was bigger in the weld than in the structure, as seen in figs 45 and 46. If the welded structure passes this bend test, the weld can then be considered of acceptable quality.
Figure 45. Example from transverse bending test face side.
113
Figure 46. Example from transverse bending test root side.
The results of the transverse bending tests are in table 32, where OK means that the weld passed the bending test. Of all the transverse bending tests, steel G got the worst results which can be explained through a number of factors.
First of all, a thickness of 12 mm, the heat flow from the fusion line was faster than in other steels. During the solidification of the molten weld pool, the porosity could increase, and these porous areas will be the first to crack during bending tests. Additionally, dilution in fusion line could be too low for the same reasons. In steel G, all the root passes failed in the transverse bending test.
This can potentially be explained by the fact that the cooling time of the root pass without being preheated is shorter in 12 mm thick plates than in 8 mm thick plates. If there are significant thickness discrepancies, it would be possible to use a three dimensional equation, however, the differences between 8 and 12 mm thickness (d) in equation 16 is 2.25 times (d2 in the equation).
In addition to heat input, cooling time is another important component in the welding process. Cooling time is dependent factor that depends on heat input, but also plate thickness, workpiece geometry, material properties and more.
The cooling time can be calculated, using equation 9.
Equation 9 allows the cooling time to be calculated with allowance for thicker plate thickness. During the course of this research, 8 mm and 12 mm thick plates of steel displayed large differences in cooling time (fig. 46-1).
Additionally, the cooling time of the root pass of QT HSS G was short, 7 s. A
114
short cooling time can lead to brittle martensite microstructure, which also has small ductile value. This is why the root pass of QT HSS G broke in the bending test.
Figure 46-1. Cooling time t8/5 vs. plate thickness. Welding conditions are presented in Table 12.
115
Table 32. Results of tranverse bending tests. OK means acceptable test.
MATERIAL WELD
Two sets of impact tests were conducted, each set including three test pieces.
Standard SFS-EN ISO 148-1 was used and each piece was 5 x 10 x 55 mm and tested at a temperature -40 °C. A 2 mm V notch was cut into each test piece and its correct placement was ensured by etching the notch before machining. The place of Charpy-V impact test is in fig. 47, which figure clarifies the structure being tested. Dependent on welding heat input, the shape of weld
116
will curve more horizontally and it leads to different HAZs under the V-groove.
As shown in fig. 47, the test area of Charpy-V test can include some weld metal, CGHAZ, FGHAZ, ICHAZ, SCHAZ and some base metal. Between first and second HAZ is the ICCGHAZ which earlier research (Liu at al. 2007, Hamada 2003, Li et al. 2001, Lambert et al. 2000, Matsuda et al. 1995, Lee et al. 1993) has shown to be the most fracture area in the HAZ. The brittle area of ICCGHAZ is small, but in some Charpy-V tests it can be under the test notch.
Figure 47. Place of Charpy-V groove in test pieces.
In earlier studies (Wang et al. 2003, Juan et al. 2003) it was noticed that lower toughness values occur because of a wide HAZ. The lowest toughness values were in CGHAZ and if the HAZ is wide all zones will be wider and then the Charpy-V test place is more in CGHAZ and fusion line. In this present research, the same results have been observed. The overall numbers of tests were small because of the testing standards, and some exceptional results are the consequence of statistical dispersion.
Overall, the results of the impact tests were ambiguous. The test results from weld area, as seen in fig. 48 and table 33 were acceptable and these results show that undermatching weld metal has good impact ductility. This might be because the impact value was limited to 18 J for test bar 5 x 10 x 55 mm piece.
As seen in table 33 steels A, E, F and G have a few results under 18 J, however
117
the vast majority of them are close to 18 J (16 J - 17 J) and they can be considered acceptable.
Fig. 49 and table 34 display HAZ impact results. In earlier study by Shi et al.
(1998) it was concluded that the lower the weld strength mismatching, the higher the fracture toughness of the HAZ. In this study, the mismatching value was very low at 0.72. There was not a great deal of consistency in HAZ impact test results, as some steels have good values for all three heat inputs while other steels had very low values. In fig. 49 the impact test values show great divergence between different HSSs. Additionally, the test values in fig. 49 are very low. Cells highlighted in yellow in table 34 indicate that the values are under standard recommendations. For example, steel H had a value 4 J twice when heat input was 1.7 kJ/mm and had poor values ranging from 8 - 13 J at 1.0 and 1.3 kJ/mm as well. Steels A, B, D and F also exhibited low impact test values, however there is no consistency in the results according to heat input.
TMCP steels A and C have low C content. C content levels determine toughness properties in general and high C content is detrimental to toughness as Hatting and Pienaar (1998) have concluded. In this study, TMCP steels A and C have low C contents, whereas the C content in QT steels was considerable bigger. Accordingly, TMCP steels have good toughness values in the HAZ than most of the QT steels.
As Tian (1998) and Hatting and Pienaar (1998) have researched, heat input has a direct effect on impact toughness in Nb added HSSs. When using a low heat input in welding, this will increase impact toughness, while if a high heat input is used in welding it will decrease the impact toughness in the HAZ. Six of eight tested steels had Nb as an alloying element in this study and the greater heat input led to the lower impact toughness.
Ti precipitations have an impact to grain growth and they inhibit it very well.
However, if the heat input is too high on welding, then the grains grow too much which leads to the coarse structure in the HAZ and consequently destroys the welded structure. Liu and Liao (1998) researched Ti nitrides and found that
118
those nitrides inhibit grain growth especially in high temperatures. However, when the temperature is too high for an extended period of time, the Ti nitrides also dissolve in the structure and their influence diminishes. This specific influence is seen in this study when using heat inputs 1.3 and 1.7 kJ/mm, where impact ductility values have decreased and grains have grown. Only steels G and H do not have Ti as an alloy element.
As Rak et al. (1997) has concluded and is also clearly displayed in this research, the size and distribution of the Ti precipitates are important when studying the grain growth control and comparing it to the role of the chemical composition of the precipitates. It is important to keep the heat input as low as possible, because Ti precipitate dissolves in to base material at higher temperatures. When the heat input is kept low, there is no time for precipitates to dissolve and the properties of the welded structure remain satisfactory. In this research, the lowest heat input 1.0 kJ/mm gives the best results of impact ductility and strength test on the chemical composition and microstructure of HSS.
In the undermatched weld structure, local mismatch can be the reason for lowered toughness. Dilution and alloying are not evenly distributed in undermatched welds and this leads to local mismatch. This study similarly clarifies the differences between impact test values as was in the Rak et al.
(1995) study.
119
Figure 48. Impact test values to weld metal using different heat input when filler material was undermatched.
Figure 49. Impact test values to HAZ area structure using different heat input when filler material was undermatched.
28
WELD Charpy V impact test values
1.0 kJ/mm 1.3 kJ/mm 1.7 kJ/mm
HAZ Charpy V impact test values
1.0 kJ/mm 1.3 kJ/mm 1.7 kJ/mm
HEAT INPUT J
WELDS
120
In addition to previous research (Wang et al. 2003), this study confirms the influence of heat input to impact toughness in HSS welding. As the heat input grows, the deterioration of impact toughness in the HAZ of HSSs is quite clear.
In this study, steels F and H had very low HAZ area impact values.
To further bolster confidence in these impact test results, and uncover different implications, it was additionally determined to conduct CTOD tests.
121
Table 33. Impact test values from weld when filler material was undermatched.
HEAT INPUT
122
Table 34. Impact test values from HAZ when filler material was undermatched.
HEAT INPUT
123 7.9. Hardness test
The welds were also subjected to Vickers hardness tests with SFS-EN ISO 6507-1 standards. The tests were conducted on the weld and HAZ areas at 0.5
The welds were also subjected to Vickers hardness tests with SFS-EN ISO 6507-1 standards. The tests were conducted on the weld and HAZ areas at 0.5