Barclay Water Management, Inc., was approached by a pump maintenance company to help resolve a pump seal failure problem associated with a 3 year old, closed loop, hot water heating system.
Because of continued pump seal problems during this time, a new pump had been installed – its seal began to leak within 2 to 3 weeks indicating that leakage was not due to an imbalance in the pump.
Barclay received a sample of the circulating water (partial analysis in the Table below) and 2 failed seals.
|REFERENCE:||Recirculating water from Heating Loop|
Analysis of the water from the heating loop showed that it was treated with a nitrite based inhibitor to a concentration of 600 mg/L as sodium nitrite. This level of nitrite is far below the 4000 mg/L upper recommended limit above which the nitrite may damage pump seals. If the nitrite concentration is traditionally this low, then it will not affect seals such as these (graphite/ceramic).
Examining the faces of the seals under the stereo microscopic showed deep and heavy scoring of the graphite and the transfer of some graphite to the opposing ceramic seal face. This is seen the following micrographs that were taken at 4 magnifications. Micrographs No.1 and No.3 show the rings where graphite has been deposited onto the white ceramic while No.2 and No.4 shows the surface of the graphite part of the seal and the heavy grooving where graphite has been torn out of the seal face. This is usually caused by the presence of small pieces of hard grit in the circulating water becoming trapped between the faces of the seals.
The water sample contained a small amount of a deposit that was black in color and highly magnetic indicating that it has a high iron component which was confirmed by qualitative analysis. The material was so small in size that a compound microscope was required to resolve its character. The following photomicrographs show 4 typical fields of view taken at 800 magnifications. The white amorphous material proved to have a soft texture and be rich in bacteria – this material will not affect the pump seals but does indicate that there may be other problems with system treatment. The black spheres proved to be small bits of welding slag from system construction/maintenance work. The welding slag is composed of iron and is, as a result, quite hard in texture. The slag seen was quite small in size ranging from 12.5 to 50 microns in diameter and is likely the cause of the deterioration of the seals submitted as the slag is small enough to get between the seals’ opposing faces and hard enough to get stuck in the graphite and score its surface leading to the grooving noted above.
THE PROBLEM: A rural hospital was having corrosion problems in its domestic, hot water piping. The water supply was from a well that was treated with liquid sodium hypochlorite as the disinfectant. We were informed that the well water has between 0.7 to 1.0 mg/L of free chlorine in the recirculating domestic hot water after treatment.
EQUIPMENT INSPECTED: Section of copper pipe removed from the Domestic Hot Water system which is approximately 1.7 inches in length (see photo No. 1). Its outer diameter (OD) was measured at 2.12 to 2.13 inches. The wall thickness measured at the thickest area was 0.067 to 0.070 inches. This corresponds to 2 inch, Type L copper tubing which has a manufactured OD of 2.125 inches and a wall thickness of 0.070”. The corroded areas of the pipe were between 0.022 and 0.45 inches in thickness but primarily under 0.040 inches.
GENERAL CONDITIONS/COMMENTS: A typical chemical analysis of the well water is presented in the Table on the left, below. Calculation of its Langelier Saturation Index (LSI) at 60°F using this data yields a value of -1.58 which indicates highly corrosive water (see the LSI Interpretation Table below right). The LSI does not take into consideration the potentially corrosive effects of the chlorine which may add substantially to the overall corrosivity of the water. A plateau on the inside of the pipe of about 1 inch in width (photograph No. 2) showed minimal corrosion. This “island” was probably on the top of the pipe where an air space existed, so little or no contact with the flowing water occurred. The edges of this area shows serious undercutting (photograph No. 3) which is a typical feature where the corrosion is a process of direct dissolution by a flowing, corrosive fluid. The heavily corroded areas had a visible “horseshoe” wave pattern (photo No. 4, green arrows) on much of the surface which is a typical feature where the corrosion has been strongly influenced by water flow. The wall thicknesses in these areas of heavy corrosion were between 0.022 and 0.045 inches which represents wall losses between 36% and 69% based upon the manufactured specifications. This corrosion is due to the effects of flowing naturally corrosive water which may have been made more corrosive towards copper by the presence of a strong oxidizing agent in the form of free chlorine which aggressively attacks the passive oxide films normally present on copper surfaces in aqueous solutions.
TYPICAL WELL WATER ANALYSIS LSI INTERPRETATION
No. 1 OUTSIDE VIEW No. 2 UNCORRODED “PLATEAU”
No. 3 UNDERCUTTING OF “ISLAND” No. 4 “HORSESHOE” PATTERNS
|THE PROBLEM::||FAILURE OF A COOLING TOWER MAKE-UP WATER LINE|
EQUIPMENT INSPECTED: Section of pipe from a Make-up Water Line to a Cooling Tower reported to have been in service for 10 years. The section was approximately 11 inches in length. The outer diameter was a quite uniform 4.5 inches and the wall thickness was an average 0.175 inches which corresponds very well with Schedule 30, 4 inch Nominal Steel Pipe which has a manufactured outer diameter of 4.5 inches and a wall thickness of 0.188 inches. The pipe is of the rolled and welded type.
TREATMENT HISTORY: The water carried by this pipe is potable water treated by the Municipal Water Department with monochloramine as the disinfectant CO2 / Na2CO3 to raise the pH and make it less corrosive.
GENERAL CONDITIONS/COMMENTS: The entire inner surface of the pipe was covered with elongated iron oxide tubercles which resulted in significant occlusion of its water side surfaces (photograph No. 2, following page). The elongation of the tubercles and their hanging orientation indicates that the section of pipe had a vertical orientation and that the pipe was completely filled with water that was stagnant or flowing slowly for significant periods of time which has caused most of the ferric ions leaching from the steel surface to precipitate as ferrous iron in a very localized fashion instead of being swept away by flowing water. When iron combines with oxygen, the resultant crystal structure is approximately 40 times its original volume such that a small amount of corrosion will be seen as a very large tubercle unless the products of corrosion are swept away by water flow. A portion of the inner surface of the pipe was cleaned of these tubercles to the underlying steel to allow close examination of the surface and measurements of wall thickness to be made. No deep pits or areas of localized corrosion were observed. A series of measurements resulted in an average wall thickness of 0.175 inches. The average wall loss would be: 0.188 (manufactured) – 0.175 (actual) = 0.013 inches or 13 Mils. Assuming the stated 10 year length of service, this results in an average corrosion rate of 1.3 Mils Per Year (MPY). This would be considered an excellent corrosion rate and would predict an expected life span of 144 years for total penetration. The standard deviation of 20 measurements of wall thickness was a low 0.0043 inches (2.5% RSD) which shows that there were no localized areas of rapid corrosion. The thinnest measured wall thickness of 0.168 inches would result in a corrosion rate of only 2.0 MPY which would predict an expected lifespan of 94 years for full penetration as a worst case.
The pipe section had a circle drawn on the outside where a failure had occurred (photograph No. 1, Red Arrow). The failure was located along the welded seam of the pipe. After cleaning, the seam along the inside of the pipe was seen to be completely devoid of any traces of weld (photograph No. 3, Blue Arrow). Along the very clean, triangular shaped groove (photograph No. 1 and No. 3, Blue Arrows) where the pipe should have been joined, the wall thickness was a maximum of 0.095 inches which represents a loss of about 50%. The void was deeper in places but could not be accurately measured. Pipe failure clearly occurred in this seam, from the water side outward. In our experience, we have never seen a weld completely corrode away with so little loss of the surrounding metal. Our conclusion is that the pipe was not properly welded and that the failure of the pipe was accelerated by this manufacturing defect.
No. 1 OUTSIDE OF PIPE
RED ARROW – Original arrow and circle showing spot of failure
BLUE ARROW – Groove on inside of pipe where weld seem should be
No. 2 INSIDE OF PIPE No. 3 INNER SEAM AFTER CLEANING
Continuous plugging of a strainer associated with a cooling tower. There was a great concern that the system was “sliming up”.
REFERENCE: COOLING TOWER STRAINER SAMPLE
Total Bacteria, CFU/mL 1,000,000
Sulfate Reducing Bacteria, CFU/mL >100,000
The sample of material that had been taken from a strainer associated with a cooling tower was found to be rich in rich in bacteria as evidenced by the Total Bacteria and Sulfate Reducing Bacteria test results given in the Table above. It should be noted that these bacteria counts were taken using a technique that measures total bacterial activity and is not from plating techniques as we are dealing with sediment and not an even suspension of the bacteria in a water sample. It is not unusual to find large populations of microorganisms associated with sediment/sludge from cooling tower basins. That being said, Sulfate Reducing Bacteria are obligate anaerobes and are often associated with Microbiologically Influenced Corrosion, an indication that control has been compromised.
In order to determine more of the nature of the sediment, aliquots were removed for examination under the compound microscope. It was immediately seen that much of the sediment was in the form of windblown debris that included seed parachutes, pollen grains and insect parts. Higher organisms including rotifers were also seen, further indication of the large population of bacteria. Some mineral scale was present.
The following photomicrographs were taken under dark phase optics and show typical fields of view. Photograph 1 was taken at 400 magnifications. This view includes pollen (arrows) along with general biofilm. Photograph 2 includes fibers from terrestrial plants, seed parachutes and other fibers that are of man-made origin (colored blue and pink in this view) which were plentiful in the strainer material. Photograph 3 shows an insect wing, and Photograph 4 displays a region of biofilm that has plant fibers incorporated within it.
Photograph 1. Photograph 2.
Photograph 3. Photograph 4.