This paper stresses the need for a reliable underwater "wet" welding process and points out the value of fundamental research in this area. Various heat transfer mechanisms are sited as being ultimately responsible for metallurgical structure and residual stress. A finite element heat transfer model including radiation, boiling and the arc bubble mechanism is described. Results are evaluated in relation to individual phenomena and areas for future research are defined.
The need for a reliable underwater joining process is very real in today's offshore technology. Although underwater chamber welding "in the dry" has proven to be a partial solution to the problem, this process is limited to the welding of regular shapes, such as pipelines and is very expensive. For more general underwater repairs and fabrication chambers are impractical and welding "in the wet" is desirable. The only other alternatives for such work are prefabrication and dry-docking, but complete prefabrication is not always possible and dry-docking is costly. Unfortunately the "state-of-threat" in underwater wet welding has advanced very little since World War II and for this reason its use has been limited to temporary repair and salvage work. The standard U. S. Navy rule of thumb for wet underwater welds predicts 80% strength and 50% ductility when compared to similar surface welds. Before the use of underwater welding can be expanded to permanent repairs and structural fabrication present methods must be modified or radically changed. There are two basic approaches to select from when solving such a problem:
Conduct an empirical study in which proposed processes might be tested, optimized, and evaluated. This first requires a proposed process to test.
Conduct a fundamental study in which the various arc, heat, metallurgical and physical phenomena may be analyzed and modeled. An understanding of these phenomena could then be used in more rational selection and development of new processes. We have chosen to undertake the second of these approaches combined with an empirical study sufficient to substantiate our mathematical models.
Such work is particularly compatible to the university whereas development of hardware and technique might best be done in industry.