Water Leaks (Weeping) Staining Investigation

Background of Inspection

On 28 May 2002 this consultant visited an office building in Pennsylvania to inspect the condition of the exterior columns of the building. There had been a report of water leaks (weeping) staining some areas of the COR-TEN steel that had been used to construct the columns. The leaks that had been reported were in the vicinity of the welds that join vertical sections of the columns, at every third floor, where spandrel beams join the columns to the building. At these locations the plates that form the box section of the columns are joined to horizontal sections of plate. The vertical flange-to-web weld joints are nominally through-section (full penetration) welds, and were shop welded using a novel j-groove technique. The box beams, once assembled, were presumably field welded to the horizontal plates or diaphragms. The welds are at both the tops and bottoms of the horizontal plates. At the exteriors of the joints the welds are approximately 1.5-2 inches in width. There is no indication that these welds were intended to be full penetration welds.

The columns are filled with an aqueous solution of 38% potassium carbonate for protection against freezing. The solution also contains 4000 ppm potassium nitrite as a corrosion inhibitor. The purpose of the solution is to dissipate heat in the event of a fire in the building. The columns are divided into four vertical zones with solid diaphragms at the 16th, 34th and 50th floors. The diaphragms between the solid diaphragms each have a four inch hole to allow the passage of the solution. Except for occasional additions of solution to correct for evaporation, the solution remains in the columns. There is no record of testing of the bulk solution (~400,000 gallons) to ensure that the chemistry has been unchanged since the original installation.

During the visit the condition of the columns and the steps that had been taken to identify the scope of the problem were discussed with the Engineering Manger and with the Maintenance Engineer. This consultant was advised that the building had primarily been inspected from ground level with binoculars, and that approximately 20 locations were identified as possible "weepers". One of the leaking joints, on the 16th floor had been repaired, but the joint had begun leaking again shortly after the repair.

Inspection by Consultant

Leaking joints were inspected at the 62nd floor and at the 16th floor. The inspection at the 62nd floor was performed from inside the building. This was a visual observation through the windows of a former restaurant. This observation indicated staining of the vertical plates of the column. The staining originated at the bottom, or underside, weld joining a section of the vertical box beam and the diaphragm. The staining occurs because CORTEN steel is not appreciably resistant to liquid phase water, and the surface of the steel is being converted to hematite, or conventional rust. The staining is obviously caused by water seeping through the weld and/or the heat affected zone adjacent to the weld.

At the 16th floor, the corrosion damage could be observed more closely by opening two windows in a conference room immediately adjacent to the leaking column. At this location the damage was similar to that observed on the 62nd floor, except that the leak originated at the top weld. An attempt had been made to repair the leak in this area, by grinding out some of the weld metal, and re-welding the joint. This repair was not successful and the joint continues to leak. The joint was clearly wet at the time of this inspection.

Consultant Conclusions

It is this consultant's opinion that the damage to the vertical columns is result of corrosion from within the columns. It is probable that the field welds are not full penetration welds, and that crevices, or narrow slots, exist between the bases of some of the vertical column sections and the horizontal diaphragms. It is likely that the crevice corrosion initiated shortly after the building was erected and is only now compromising the integrity of the welds. Even if the majority of the solution has retained an appropriate level of corrosion inhibitor, it is well known that crevice corrosion can propagate because mass transport in the solution is limited in very narrow slots. The extent of damage cannot be ascertained at this time. The degree to which leaking will occur will be a function of the weld integrity, the nature of the crevice (narrow crevices re more damaging than more open crevices), and the present chemistry of the solution. In any event, the damage cannot be ascribed to single event and in this consultant's opinion, has been progressing over tens of years, and can be linked to the original design and construction of the columns.

To see the resume of the expert associated with this case study, see the link below.