Hollow bead - a question of technique?

TWI Bulletin, May/June 1993

Mike Wright is a Senior Research Metallurgist in TWI's Materials Department. He joined TWI in 1989 after completing post-graduate studies on the fatigue properties of high specific modulus aluminium based materials at Birmingham University. His work at TWI has been concerned with the weldability of ferritic steels. He has a particular interest in hydrogen cracking.

Hollow bead defects in pipeline are not common, but when exposed, can be costly to eradicate. Little quantifiable research on the phenomenon exists. Indeed most of what is known is founded in rumour, conjecture and welding folklore. Having examined the literature available on linear porosity and industry's experience of it Mike Wright believes that welding current and travel speed are largely, but not totally, responsible.

The defect known as 'hollow bead' is generally associated with stovepipe welding of pipelines ( Fig.1) where cellulosic consumables are used in the vertical-down position. As the name suggests, hollow bead is an elongated void found within the root bead and can extend for some length along the weld. It is also known as piping or linear porosity ( Fig.2). The extent to which this type of defect compromises the integrity of the weld can be uncertain, as in many instances it is located entirely within the reinforcement of the root bead and can have a relatively benign circular cross-section. Nevertheless, its presence is readily detected by radiography and specifications such as API 1104 require that it be repaired even if it is present in only small amounts.

The cost of such repairs during pipeline welding is considerable, so an understanding of the mechanisms behind the formation of hollow bead and, most importantly, how to prevent it are of great interest. Unfortunately, although a serious problem when it does occur, hollow bead tends only to manifest itself occasionally and often in situations where production pressures or geographic location preclude any investigation. Thus, emphasis is currently placed on cure rather than prevention.

Perhaps, because of this, most reports of hollow bead are anecdotal leading to confusion over the real causes of the problem. In general, hollow bead is accepted to be process or operator-induced but other factors are frequently cited, particularly parent material composition. In order to establish the validity of any links between parent material composition and hollow bead, TWI has undertaken a review of published literature on the subject. In addition, the experience and opinions of welding engineers, consumable manufacturers and pipeline operators, with knowledge of stovepipe welding have been sought. A list of companies contributing to this review is given in Table 1. The results of the review are outlined in this article.

Table 1 Companies contacted as part of this survey

Published work on hollow bead

The confusion over the causes of hollow bead can in part be explained by the scarcity of published research on the subject. It appears that the only thorough investigation of the phenomenon was carried out by Barkow in the early 1970s. The results were presented at the API Pipeline Conference in Dallas 1973, and in an abridged form in the Oil and Gas Journal of July 1973. ^[1] This work was performed in response to a reported increase in the incidence of hollow bead with the introduction at that time, of fully killed, high strength, high toughness linepipe steels. It had been suggested that the higher silicon contents (up to 0.40%) of these steels was responsible for the problem. This was despite the fact that hollow bead had previously been reported in semi-killed steels.

As a result of his work on API 5LX65 steels containing up to 0.35% silicon, Barkow concluded that welding speed was the real cause of the problem and that a welder could correct the condition merely by changing his technique. In particular, the problem could be avoided by use of a lower travel speed. The explanation for this was that, when using a high welding speed, less metal is actually deposited. Hence, the molten metal available is not sufficient to fill the entire root gap, and, as solidification is from the outside in, voids are created at the centre of the weld nugget.

If, on the other hand, welding speed is lower, more metal is deposited and more heat introduced, keeping the weld pool molten longer and giving time to completely fill the root gap. This hypothesis of Barkow's does not fully explain the rounded nature of hollow bead which suggests some contribution from entrapped gas, but any such contribution would probably also be exacerbated by excessive travel speeds. ^[2]

Although Barkow concluded that welder technique was the real cause of the problem, he did recognise that the higher silicon contents, of the then new, fully-killed pipeline steels would result in slightly more fluid weld pools. This in turn might cause welders unwittingly to use excessive travel speeds in order to compensate, rather than adjust other variables. It was implied that, once a welder was used to these steels and aware of the need to limit travel speed, there would no longer be a problem.

It is worth mentioning that during an extensive investigation of hydrogen cracking of cellulosic welds in API 5LX60 pipe, carried out by TWI in 1969 ^[3] no problems with hollow bead were reported. In this case the silicon content of the pipe was 0.38%.

Subsequent publications referring to hollow bead are rare and either merely refer back to the work by Barkow ^[1] or are unsupported by experimental data. An example of the former is an article ^[4] published by the Welding Research Council (WRC) describing the nine most common pipeline weld defects and their significance. This article (referring to Barkow's work) attributes the defect to excessive welding speed and current.

Hollow bead is mentioned in an article by Düren and Müsch describing defects arising during site welding with cellulosic electrodes. ^[5] In this piece hollow bead is listed among other defects associated with the stovepipe welding technique which have become much less common since equipment and welder training were improved.

An alternative source of information on weld defects is welding guidelines produced by consumable manufacturers. The Lincoln Electric Company produces a booklet ^[6] titled 'Welding pressure pipelines and piping systems' which contains a root pass trouble shooting guide and a section describing special techniques for X60, X65 and X70 pipe, both of which mention hollow bead. In these, it is recommended that DC(-) be used to avoid hollow bead and that, if it does occur, then current should be reduced and the gap increased. In addition to these recommendations, a warning is given that hollow bead generally occurs with thin wall pipe and pipe steels containing over 0.1% silicon, which is described as high. This statement is no longer valid as silicon contents are now typically 0.25 to 0.35% ^[7] and levels in excess of 0.40% were sometimes encountered in the 1970s, see Table 2.

Table 2 Development of pipeline steels used in the British Gas Transmission System (from Ref.11)

Cellulosic consumable manufacturer Bohler also includes a trouble shooting section in its pipe welding consumable guide. ^[9] In this, hollow bead is linked with narrow root gaps and high aluminium content (greater than 0.04%) in parent material. The link with high aluminium is apparently based on the results of an internal research project ^[10] performed in 1972.

Industrial experience of hollow bead

The survey of welding engineers and others with experience of hollow bead problems revealed a consensus of opinion that hollow bead is a welder-induced defect, the primary cause being welding at too great a travel speed or too high a current. Two contacts mentioned that it was usually restricted to individual welders. Further discussion revealed, however, that it was recognised that factors beyond the control of a welder might influence the tendency to form hollow bead although not to the same extent to which parent material affects hydrogen cracking susceptibility. ^[11]

Of those questioned, four contacts thought that parent material chemical composition might have a bearing on the problem. Interestingly, only two people contacted as part of this survey mentioned the possibility of silicon content in parent material being a factor. Three out of four mentioned a trend to form hollow bead with very low sulphur steels and two mentioned a similar trend in steels with high aluminium content. It should be noted however, that these were recognised as minor factors compared with welding technique and control of speed and current.

A factor identified by nearly all those questioned, which also has a bearing on welding technique, was poor handling characteristics of electrodes, particularly due to inadequate coating moisture content. A summary of response to the survey is given in Table 3.

Table 3 Breakdown of responses to the telephone survey of industrial experience of hollow bead

Similar information was forthcoming from a member involved in general fabrication, who had observed hollow bead type defects in welds deposited in the vertical-down position. In this case, the type of consumable was identified as having a strong influence, presumably because of variations in handling characteristics.

TWI Experience

It was mentioned earlier that hollow bead is a fairly uncommon occurrence. Nonetheless, it is somewhat surprising that, to the author's knowledge, TWI only once has been requested to investigate an incidence of hollow bead. ^[12] This was in the late 1960s when a fabricator reported difficulties in eliminating root bead porosity which was resulting in a repair rate of 70%. The investigation concluded that the main defect giving rise to repairs was linear root bead porosity (hollow bead) but that shrinkage porosity was also present. The cause of the defect was not positively identified but the chemistry of the pipe (API 5LX52 of 0.33%Si) was discounted as having any influence. It was suggested that faulty welding technique and/or defective electrodes were the most likely cause. The report recommended in that situation that, in order to eliminate the problem, welds should be made vertically-up. It should be noted however, that this could not usually be recommended because of the loss of productivity associated with a change from vertical-down to vertical-up.

Summary

The results of this survey indicate that hollow bead should be regarded as an operator/process induced defect which can be avoided by changing technique. The primary control should be in setting a maximum travel speed, but there may well be advantage in reducing the welding current.

The role of parent material composition, and particularly the silicon content, appears to have been over-emphasised and misunderstood. This survey indicates that it is possible to produce welds under site conditions completely free of hollow bead up to at least 0.38% silicon. The survey has also shown that hollow bead type defects can occur in welds other than pipeline stovepipe welds if conditions such as travel speed and welding position are conducive to its formation.

References