 Remember what you learned in class about sharp corners acting like stress risers? That problem doesn't end with machined components. High profile welds can act like sharp corners that create stress concentrations; and it's no surprise to see them fail in fatigue at the toe of the weld. In this part of the world, when a stainless steel weld fails in fatigue, it always starts along a high profile weld toe.
Are stainless steels notch sensitive, or have we discounted the importance of smooth weld profiles? The data tends to say that stainless steels have good notch sensitivity values. However, a high profile convex weld bead, in effect, creates a notch at the weld toes. Also, several convex high profile stringer passes create notches between each pass. The failures speak for themselves. When designing with 40 to 80 million load cycles in mind, those weld toes and notches between stringer passes have got to go. A smooth profile is important. The rest of the story ... The all stainless steel vessel (Fig. 1 left) was about to fall down. Could anyone tell? The answer was no. It was by a fluke of nature that the problem was noticed in the first place. The problem had been concealed by insulation for an unknown length of time. The vessel was 12-years old. It was by sheer luck and a keen eye that the problem was detected in the first place. It was serious enough to make one pause for a moment and think about the ramifications of hidden damage. Damage that had never been detected by any normal inspection program because of the very fact that the insulation had hidden it from view. The bottom half of the vessel was in the shape of a cone. That cone was supported by a skirt whose diameter was smaller than the greatest diameter of the cone (for a close up of the skirt point to it in Fig. 1 above). Hence, the very nature of the design caused a wedging action between the cone and the skirt support. The skirt supported the entire vessel close to the transition between the cone and the vertical sidewall of the vessel, hereafter referred to as the "knuckle." The skirt was welded to the cone (Fig. 2, lower left) at this location. The cause of failure was shearing of the attachment weld at the cone-side weld toe. Wherever it did not shear, it was because of a lack of fusion on the skirt-side weld foot. For a close up of the weld point to the highlighted region in Fig. 2 lower left. |
|
 |
 |
| When the suspicious cracking was noticed, the insulation was removed around the entire knuckle area. It revealed that there was only 24" of weld still supporting the vessel (Fig. 2, upper left). The vessel had flared the skirt outwards breaking the stiffening ring welds and causing the stiffening ring to fall off. In so doing, the vessel dropped down by as much as 3" into the skirt (Fig. 3 upper right). There were only two other locations where the vessel was supported. In one location (Fig. 4 lower left) the vessel had fallen until it contacted the stiffening ring which now supported the vessel (the ring hadn't broken away like the ring in Fig. 3 upper right). In the other support location (Fig. 5 lower right) the knuckle had grabbed the skirt on its way downwards and caused it to crimp. If the skirt had ever developed a vertical crack at for instance, the crimp, the vessel would have sliced through it and went straight to the ground. | |
  |
|
 Taking a closer look at the attachment weld between the skirt and cone (click on the highlighted weld in Fig. 2 above) the angle that the weld face made with the cone was close to 90 degrees. This sharp angle created a stress riser. Couple this stress with the shearing action of the cone as it wedged into the skirt by shear gravity, and this weld was destined to fail.
The repair consisted of improving the attachment weld profile by grinding it smooth and increasing the transition angle between weld and cone (Fig. 6 right). For a close up view of the weld click on it in Fig. 6. In addition, another stiffening ring was added below the attachment weld, on the skirt support. This additional ring would provide lateral stability of the skirt. To improve upon the weld quality, an NDE contractor and a CWI was brought in to radiograph and supervise the welding. there was evidence of lack of fusion in a lot of the old weld (Fig. 2 highlighted region). The attention to weld quality and weld profile detail is covered in the August MACHINE DESIGN article. Radiography, ultrasonic shear wave, and dye penetrant methods were utilized to evaluate weld quality. |
|
| All Pictures and Text Copyright © 2000 - 2016 Contact Mr. Adler Adler Engineering LLC of Wyoming USA |
Great care has been taken in the compilation of this article. However, no warranty, expressed or implied, including without limitation, warranties of merchantability or fitness for a particular purpose, are given in connection with this article or any article archived on this website. Although this information is believed to be accurate by the author, the author cannot guarantee favorable results will be obtained from the use of this article alone. This article is intended for use by persons at their sole discretion and risk. Since the conditions of product or material use are outside of the author's control, the author assumes no liability or obligation in connection with any use of this information. The author is not liable for special, indirect or consequential damages resulting from the use of this material.
No part of this article or any article archived in this website, or any part thereof, may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning or otherwise, without the prior written permission of the copyright holder R. H. Adler. Nothing contained in this article or any article archived in this website shall be construed as a grant of any right of manufacture, sale, use, or reproduction, in connection with any method, process, apparatus, product, composition, or system, whether or not covered by letters of patent, copyright, or trademark, and nothing contained in this article or any article archived in this website, shall be construed as a defense against any alleged infringement of letters of patent, copyright, or trademark, or as a defense against any liability for such infringement.