I conducted a tensile test to analyze the strength of a welded joint and to implement an FEM model. Due to the high hardness of the base material, slippage between the grip and the base material was anticipated, so an additional jig was fabricated for the test. The specimen consisted of two base materials, each 50 mm wide and 120 mm long, connected by welding, with a total length of 240 mm. Since the strength of the welded joint was expected to be weaker and its area smaller compared to the base material, the deformation was anticipated to be localized to the welded joint.
In the test results, the specimen reached the fracture point after a total deformation of approximately 6 mm. However, as expected, the actual specimen exhibited deformation localized to the welded joint, with less than 1 mm of deformation (a difference of about 5 mm). Initially, I thought this error was caused by slippage between the grip and the jig, but after reading Han Lu's question, I reconsidered.
Han Lu's Question
:https://www.researchgate.net/post/How_to_calculate_Youngs_modulus_in_uniaxial_tensile_tests_conducted_with_or_without_an_extensometer
1) If there was no slippage between the grip and the jig, is it reasonable to think that the difference in deformation between the specimen and the test data(almost 5 mm) is due to the deformation of the machine?
2) If the deformation in the test data includes the machine deformation and thus cannot be trusted, can the maximum load be considered a reliable result?
3) Given that the length of the welded joint is quite short, should I consider using DIC or an extensometer to obtain accurate data?
In conclusion, the test was conducted to implement a model for FEM analysis, but I would like to know if such data can be used to create the model.
Hi Minseong,
The determination of the material strength using the test you have described might be possible, but it is a bit prone to error for the following reasons.
- The stress will no be uniform across the weld bead. This means that a simple load over area calculation will only provide the average stress, and there may well be regions where the stress is higher.
- The stress state will not be uniaxial in the weld bead. This means that the effect of the hydrostatic stress on the material strength measured will not be known.
- The weld beads may contain residual stresses and these are likely to be tensile (there is a saying that the the residual stress in the last part to cool is tensile). The residual stresses may have been mechanically relieved at the point when the welds break, but only if they break in a ductile manner. Otherwise, they may affect the load to break.
- (Are the before and after pictures of the same specimen? It seems that the after picture has more welding spatter than the before picture.)
- The "After gap" seems bigger on the left than on the right. Does this mean that the welds were not loaded equally?
- There are standards for the measurement of weld metal properties (e.g. see ISO 5178:2019).
Regards,
Simon
Simon Smith
Hello Simon,
First, this test was conducted to measure the tack weld strength of the welded joint. Therefore, I looked for regulations regarding the testing of general welding metals and welds, but since there were no specific regulations for tack welding strength, the test was performed arbitrarily.
The answers to your questions are as follows:
(Are the before and after pictures of the same specimen? It seems that the after picture has more welding spatter than the before picture.)
: Although the specimens were not identical, all five tests showed the same fracture pattern. You can consider this as an example of the fracture pattern.
The "After gap" seems bigger on the left than on the right. Does this mean that the welds were not loaded equally?
: The "After gap" photo shows the broken specimen reassembled, which may appear misaligned because the fracture surfaces do not match perfectly. There was minimal gap difference between the left and right sides.
Thank you for your response. Considering your points, we can conduct the next experiment more accurately.
Additionally, my question was about the significant discrepancy between the deformation of the specimen and the deformation measured during the tensile test. After consulting with local PhDs and industry experts in tensile testing machines, I was informed that the displacement of the tensile testing machine is measured at the crosshead, which means the machine itself can deform, causing this discrepancy. To obtain accurate values, we need to use an extensometer or strain gauge.
The experiment was conducted using a mechanical tensile testing machine with a capacity of 300 kN (servo motor type). Despite the fracture load being around 70 kN, which is considerably lower than the capacity, such discrepancies still occurred, which I find hard to understand. For the next experiment, I plan to use an extensometer or strain gauge.
Best Regards,
Minseong Cho
Hi Minseong,
You might be interested to know that research has shown correlations between hardness and tensile strength. The following article says that "multiplying the Vickers hardness number of a carbon steel by 3.3 will give the approximate ultimate tensile strength in MPa".
https://www.twi-global.com/technical-knowledge/job-knowledge/hardness-testing-part-2-075
The article is on the TWI web site. I worked at TWI for over 20 years.
Would you mind saying why you are interested in the strength of tack welds?
Regards,
Simon
I assume that the pictures showing are thick plates. but your test for a thin sheet. so, please modify them based on the machine or increase the pressure on jig
Dear Simon Smith ,
I am interested in tack weld strength because there is a process of performing tack welding and lifting for large structures in order to complete the final welding. To ensure structural stability during the lifting process, I need to confirm the adequacy of the tack welds. Ultimately, I aim to perform CAE analysis based on the data obtained from the tests.
Hi Minseong,
Thank you for your reply with the explanation of the need for the strength of a tack weld.
I have worked on the structural analysis of welded joints for many years, including the use of FEA.
Will your final CAE model be made from brick elements? If yes, will you have a fine mesh that is able to predict the stress distribution in the fillet weld? And, if so, will you predict the tack failure when the peak stress (or strain) is equal to a critical value. Or, will you determine the average stress in the tack weld and compare it to the measured strength in the tests?
Also, I assume that some tack welds will experience both direct and shear loads. It seems that your testing only applies a direct load.
You might find that previous work on fillet welds would be useful, for example:
Effect of loading angle on the behaviour of fillet welds
https://www.aisc.org/globalassets/aisc/research-library/effect-of-loading-angle-on-the-behavior-of-fillet-welds.pdf
And
Fillet and PJP welds
https://www.aisc.org/technical-resources/research/researchlibrary/fillet-and-pjp-welds/
Regards,
Simon
- Badges
- Science topic
- Similar topics
- Papers