Don’t Ignore Small Failures

While a power plant’s fractured oil plug seemed like a small failure, the root cause suggested larger, more costly incidents could follow.

By Randall Noon, P.E.

The following case study involves a fractured oil plug in a power plant. In itself, a fractured oil plug isn’t a significant failure in a large industrial facility. However, even back-of-the-envelope analyses of seemingly small failure incidents can reveal larger patterns that should be addressed.

INITIAL FAILURE STORY

The oil plug shown in the above photograph had been removed from a 300-hp vertically mounted electric motor. That motor drives a mixed-flow, single-stage vertical pump in a power plant. The pump feeds cooling water from a river to various heat exchangers. The plug is from the lower oil reservoir that services the lower motor bearing. An upper reservoir services the motor’s upper bearing.

The fracture was discovered when the plug was removed to collect a lubricant sample for periodic analysis. Leakage from this plug had been occasionally reported in the past. Each time a leak was noted, a mechanic was dispatched to fix it.

No other information was supplied at the time the photograph was provided for review. What could the image possibly tell us?

OBSERVATIONS

The plug’s threads (standard right-hand type) are undamaged. The head of the plug has tool marks on the edge of the flat, i.e., in the top left in the photograph. These marks are consistent with those made when a tool slips off the head. In this case, they appear on the side of the head typically engaged when the plug is being loosened.

The edge between the two visible flats is rounded off. This typically results when a tool slips on the head. There are tool marks opposite the aforementioned ones, near the thumb of the person holding the plug. These marks are serrated, similar to those that would be made by a tool with gripping teeth.

The plug is equipped with an O-ring, i.e., gasket. Instead of being round, the gasket profile is misshapen. A ridge down the middle of the gasket indicates it had been crushed by tightening and had not elastically recovered its original round shape.

Note that the crack in the plug begins at a thread root. This crack is within the smallest diameter of the plug’s shaft and has a barber-pole-type fracture angle of about 15 deg. The crack also passes through a hole drilled through the smallest diameter of the plug. Superimposing the two sides of the hole bisected by the fracture, one can see that the hole was slightly distended in the area where the crack passed through it. This suggested the plug was subjected to relatively high levels of tensile stress.

ANALYSIS

The crack started in the weakest part of the plug with respect to axial tensile stress: at the root of a thread where the shaft diameter is smallest and where there is a 3X stress riser.

The 15-deg. angle of the crack indicates that not only was the area of the fracture in tensile stress, but that torsional stress was also present. The angle of fracture provides a way of determining the ratio of tensile stress to torsional stress that was present when the plug fractured. Recalling a little mechanics of materials, i.e., Mohr’s Circle of Stress equations, assuming there was tensile stress due to tightening, as well as shear stress due to tightening or loosening of the plug, then the following holds true:

tan(2q) =  (2t)/s

where q = fracture angle of the crack

t = torsional stress

s = axial tensile stress

In this case, the tangent of 2 x 15 deg. (or 30 deg.) is 0.577. So, when the crack occurred, the torsional stress being applied was equal to about 28.8% of the tensile stress that was present. Substituting this data into the maximum principle stress equation for a two-dimensional case, the maximum principle stress that existed when the plug fractured is as follows:

s_max =  s/2 + [(s /2)² +  t²]^1/2

s_max  = 1.077 s

The plug clearly fractured when it reached the limit of its ability to resist internal stress. Thus, s_max (above) is equal to the ultimate strength of the material. The preceding equation also tells us that the oil plug was tightened such that its axial tensile stress was about 93% of its ultimate material strength.

This amount of tensile stress is unnecessary for a simple oil plug. The plug is not a structural bolt that has to be stretched nearly to its breaking point.  It merely requires enough tension to compress the rubber O-ring or gasket sufficiently to prevent oil leakage from a low-pressure oil system.

In applying torsion to the plug, shear stress was added to the axial tensile stress. The additional shear stress caused the principle stress angle to shift about 15 deg. The combined tensile and shear stresses then exceeded the ultimate strength of the material. In short, the oil plug was tightened until it broke.

WHY?

The motor-maintenance manual recommends replacing the gasket after the oil plug has been tightened twice. This is a reasonable recommendation since the elastomer in the gasket, given time, degree of compression, and ambient operating temperature, won’t rebound elastically to its original shape. With each cycle of compression and release, the gasket becomes misshapen just a little bit more.

Review of the motor’s maintenance history revealed that no replacement gasket had been released from inventory for this specific work task in several years. Each time an oil leak was reported, the plug was simply tightened until it the leakage stopped.

In failure reports from other power plants that had experienced similarly fractured oil plugs in similar machines, some authors blamed the original material in the oil plug for being too weak. Personnel at those sites then changed the original material specification for the plug to a stronger material. They apparently reasoned a stronger material would allow them to tighten the oil plug as needed without fear of fracturing it.

If the primary concern of those power plants was the oil plug itself, that rationale would hold true. The additional compression on the gasket afforded by greater tightening of the oil plug would allow the gasket to be used several additional times before leakage occurred or the plug fractured. Unfortunately, by electing to use oil plugs made of a stronger material, site personnel risked damaging the pump-motor housings. The economics don’t make sense, considering the modest cost of new gaskets versus the cost of a re-designed plug or replacing a cracked pump-motor housing.

LESSONS TO BE LEARNED

Several lessons can be learned from this case study.

Gaskets are cheap. Replacing one every time the oil plug was removed would have been better than replacing it every other time, given the fact that doing so could reduce the amount of tightening needed to compress the gasket and stop leakage.

Changing the material specification to make the oil plug stronger would actually introduce unnecessary levels of stress into the pump-motor housing. In such cases, the risk of cracking the housing will increase. Compare the cost of a gasket to that of re-designing the oil plug, or replacing a cracked 300-hp pump housing. It’s significantly less—very much so.

Equipment maintenance manuals are valuable resources, but only if they are read. The facts in this story indicate that the power plant’s personnel, for whatever reasons, didn’t follow the motor manufacturer’s recommendations. How often was this occurring with regard to other plant assets?

In the end, the most important lesson for other plants is to ensure this scenario isn’t occurring in their operations. EP

Randall Noon is a registered professional engineer and author of several books and articles about failure analysis. He has conducted root-cause investigations for four decades, in both nuclear and non-nuclear power facilities. Contact him at
noon@carsoncomm.com.