Omer W. Blodgett, Sc.D., P.E.

How large must a weld be when no load is transferred between the two parts? If it is re-ally true that "no load" is transferred, then the welds need only be large enough to keep the parts in their proper relative positions. But if the application involves "elastic loading," achieving the required weld size may be essential. Overlooking such loading could spell disaster.

Take the example of a huge rolling mill used to reduce the thickness of steel.

It consists of a pair of large-diameter rolls, called backup rolls, and a smaller set, called working rolls. The back-up rolls drive the working rolls and provide mechanical support to keep the smaller working rolls from deflecting. The steel passes through the space between the working rolls.

To reduce the thickness of the steel being processed, the working rolls must apply high contact forces. These forces transfer from the working rolls to the back-up rolls, and from the back-up rolls into the large frame of the rolling mill. Although the contact forces are great, the massive size of the frame of the mill results in a low stress level of only 3.5 ksi.

Supporting the working rolls is an 18-in.-thick block of steel that must be joined to the frame of the rolling mill. The block simply supports the rolls, and no load transfers from the roll to the block, nor will any load transfer from the block to the frame. The design engineer has the task of deciding how large the welds need to be.

Assuming no load will be transferred, the engineer consults AWS D1.1, Table 5.8 to determine the minimum prequalified fillet weld size (in this case, ⁵ /16 in. for steel 6 in. and greater in thickness). Putting a weld of a mere ⁵ /16 in. on such massive materials seems illogical, so the engineer analyzes the problem further and considers the potential for elastic loading.

The 18-in.-thick block is 30-in. long.

Since no loads transfer from the working rolls to the block, it is exactly the same length when unloaded (for instance, when no steel is being rolled) as when loaded. In contrast, however, a 30-in. length of the mill's frame is 30-in. long when unloaded, but when steel is being rolled, the 3.5-ksi stress level causes the frame to stretch. Mathematically, the elongation calculates to 0.0035 in. Even though the block will not stretch in service, a weld attaching the block to the frame must transfer the forces resulting from this elastic loading.

Stretching the 18-in. block the same 0.0035 in. requires that the block not only elongate, but that it also bend. It turns out that the force required to do this is 48.4 kips/linear in., and the weld that attaches the block to the frame must resist that load. Using static design allowables, this requires a fillet weld with a 2 ⁵ /16-in. leg. Given that the assembly is subject to cyclic loading, the engineer increased the leg size to 6 in. to keep the stress ranges below 8 ksi.

While initially, the engineer expected that only a minimum weld size would be needed, he ultimately realized that elastic loading required a much larger weld.

In another case, unequal temperatures created an elastic loading situation. An aircraft manufacturer made a special 4-in.-thick steel platen to hold an airplane wing during machining.

The aircraft material had to be held at a temperature of 400° F during the machining operation; therefore, the platen was held at the elevated temperature as well.

Later, the manufacturer added an extension to the platen. The extension was only 1-in. thick, and there was no requirement for the portion of the wing supported by the extension to be held at the higher temperature. Therefore, only the original 4-in. platen operated at 400° F. The manufacturer estimated that the 1-in. extension would achieve a maximum temperature of 100° F.

Since no loads would transfer between the 1 and 4-in.-thick sections of steel, the manufacturer added a ³ /16-in. partial joint penetration groove weld to one side, and a ³ /16-in. fillet to the other. When the modified platen was heated for the first time, the weld tore full length.

The 300° F temperature difference between the two materials resulted in elastic loading, causing the 4-in. piece to elongate 0.072 in. over a 40-in. length.

The force required to elongate the 20-in.-long, 1-in.-thick steel to 20.036 in. was 54 kips/in. The manufacturer solved the problem with a continuous CJP groove weld. This example illustrates the significant force that can be caused by elastic loading.

It doesn't happen every day, but when elastic loading occurs, the designer must consider it to avoid unanticipated performance problems.

Omer W. Blodgett, Sc.D., P.E., senior design consultant with The Lincoln Electric Co., struck his first arc on his grandfather's welder at the age of ten. He is the author of Design of Welded Structures and Design of Weldments and an internationally recognized expert in the field of weld design. In 1999, Blodgett was named one of the "Top 125 People of the Past 125 Years" by Engineering News Record. Blodgett may be reached at (216) 383-2225.