Fatigue strength of multi string weld 1

[hpon]

Hi,

I have been asked to say something useful about the fatigue strength of a novel kind of weld and how it compares to a normal fillet weld. ..and I find it tricky.

The joint is essentially a one-sided T-joint. The weld is a mix between a butt-weld and a fillet-weld. First, a partial butt-weld is created and subsequently a fillet-like-weld is drawn on top of the first one. The loading is perpendicular to the weld.

I suppose I could analyze it using the effective-notch method and get some result. But, I'm suspect that the double-layering of the weld will reduce its fatigue strength in ways that are not captured by this analysis method.

What do you think? Is the effective-notch method applicable in this case? Is it possible to say something qualitative about the weld without a thorough analysis?

Best Regards,
hpon

 

[metengr]

hpon;
I believe in reading your post above what you are describing is a partial penetration butt weld with an external fillet (T-joint). Is this correct? Can you post a drawing of this novel kind of weld?

 

[hpon]

Metengr,

Yes, that is exactly what it is. Perhaps the word novel is an over-statement though

I have attached an illustration of the weld.

/hpon

 

[metengr]

This weld joint design will have a significant knock-down factor in reverse bending cyclic (fatigue) service. The main reason is you have a sharp notch at the weld root for the partial penetration weld. If you wanted to reduce the effects of the knock-down factor for fatigue, you should really have a full penentration butt weld followed by a small fillet.

 

[MikeMet]

Whether you make full penetration weld or not, the weak point for fatigue crack initiation is still the toe of the fillet weld, even on a small fillet weld. Studies have shown that autogenous TIG washing the toe is the most effective means of reducing the fatigue susceptibility of the weld toe. Other method is to grind or sand flapper wheel the toe to a smooth transition (radius) with the base metal to reduce the notch formed by the geometry change at the toe and / or shot peen the smoothed toe to induce compressive stresses. I have attached two generate fatigue curves showing the effect of notches, such as those at the toe of a fillet weled, on fatigue life

 

[hpon]

Thank you very much for your replies!

Is it only the general geometry that matters? Or, does the layering of the welds cause other effects such as reduced weldability for the second weld or an increase of defects?

/hpon

 

[hpon]

..and, do you think this layered weld is better or worse than a fillet weld with similar dimensions (I leave the definition of "dimensions" open for interpretation)?

Furthermore, the loading is mostly in the vertical direction and applied on the non-welded side (see attached file in third post). Therefore, I would not be too surprised if the failure mode is a root-crack.

/hpon

 

[metengr]

and, do you think this layered weld is better or worse than a fillet weld with similar dimensions (I leave the definition of "dimensions" open for interpretation

This weld will be worse than a typical fillet weld in terms of fatigue strength under your stated conditions of loading in service.

 

[hpon]

Why is that?

Regards,
hpon

 

[metengr]

You have three sites for fatigue crack initiation - the toes of the fillet weld and the weld root location.

 

[hpon]

I don't understand what you mean. A normal fillet weld have a root and two toes as well.



Regards,
hpon

 

[metengr]

hpon;
In your situation because of the direction of loading, I believe the partial penetration weld will result in increased susceptibility to fatigue crack initiation from the weld root. If this were just a normal fillet weld, the weld toes would be more susceptaible to fatigue crack initiation in comparison to the weld root.

 

[civilperson]

Layering or multiple passes done with correct welding technique should have no fatique considerations.

 

[hpon]

My thinking was that a butt weld generally is better than a fillet weld, and that the proposed weld is a mix between these two and therefore is better than a fillet and worse than a butt weld. My skepticism to this way of thinking mainly relates to the quality of the proposed weld (will there be additional defects due to the welding method?).

But you putt forth the actual geometry as a weakness compared to fillet welds. Could you expand a little on how you arrive at your conclusion? I would really appreciate that.



Regards,
hpon

 

[stanweld]

hpon,
The best T-joint design for fatigue service is a full penetration groove weld with reinforcing fillets on both sides with additional preparation per MikeMet.

Under the loading parameters described, the design proposed is not considered good with respect for fatigue. Essentially the unwelded side will lift under tension placing a high shear load at the root of the partial penetration groove weld.

 

[hpon]

Thank you for that information.

Earlier generations of the product used a fillet weld instead of the proposed kind. How do these two compare?

Regards,
hpon

 

[stanweld]

A properly sized double fillet weld of the T-Joint would be superior under the described loading parameters.

I would ask why are you making the proposed weld noting the additional weld preparation required? Have you had previous failures with a double fillet T-Joint?

 

[hpon]

Due to the design of the product, it's only possible to place a weld on the "out-side" of the joint (double fillets are not an option). And some recent changes caused production issues that eliminate the single fillet weld as an option too. The welders think that it's possible to make the proposed weld however, and hence the question popped: "will the fatigue strength remain at least as good as before?"



Regards,
hpon

 

[stanweld]

Your best bet would be to make a full penetration groove weld from one side with reinforcing fillet on the accessible side.

 

[hpon]

OK. Thank you for the advice!!

Regards,
hpon