Consider the following hypothetical.
You have installed the largest melt pump in the world, one for each of your two trains. The management team expects record production rates in the new polymer plant. The maintenance group anticipates little to no problems with these melt pumps, as the smaller-sized pumps have proven effective in other production facilities. The plant celebrates meeting design production rates. That Friday evening, you leave the plant after checking the discharge pressure and temperature of the melt pump, and all is well.
While relaxing over the weekend, you receive a call from the plant. The melt pumps have seized up. Then, the nightmare begins. You go to the plant. "All hands are on deck" to shut the plant down and pull the melt pump. Next, you meet in the plant managers' conference room to discuss the plan to get back up and running, which includes a root cause failure analysis (RCFA). You are in charge of an RCFA team chosen by management comprising representatives from maintenance, engineering, process/ controls engineering and operations. Because the pump had been running well and within the design's capability, its failure is a mystery.
In the first phase of the RCFA, the process data is evaluated. You notice the temperature of the polymer lubricated bearing was elevated just prior to the failure. This is not surprising, given the plant has started running fractional melts, which in layman's terms is a more vicious product. The rotating equipment engineers run their bearing program and determine that the unit loads were within the normal range. Polymers are typically "non-Newtonian," meaning the viscosity is a function of shear rate and temperature. However, the bearing program, which utilizes a constant viscosity, fails to address the polymer lubricated bearings. After the melt pump is removed, the metallurgical and materials group evaluates the bearings. The morphology of the failure suggests a lack of lubrication. The team discovers that the silver babbitt in the bearing had melted and froze the bearing to the shaft. You determine that the bearing was running smoothly with the same polymer until production increased.
In this problem, the RCFA revealed that the root cause was lack of lubrication. The lack of lubrication was caused by several factors, which were:
- The polymer started to break down at the higher temperatures.
- The pump suction design prevented the teeth from filling properly.
- The feed vessel level was not high enough to prevent non-condensables from getting into the polymers.
- The lube groove in the bearing did not allow for enough mass flow rate of polymer to lubricate the bearing.
To evaluate such a problem, one should take a multi-disciplined approach involving metallurgical analysis, process, controls, mechanical and operations as follows:
- The polymer found in the bearing should be analyzed for breakdown. In fact, the rheology should be determined across the entire temperature and flow range.
- Computational fluid dynamics analysis should be conducted to characterize the flow and temperature. The results of this analysis will determine the 3-D flow field including the temperature and local pressure contained within the polymer film in the bearing.
- Finite element analysis should characterize the temperature distribution and shaft deflection.
- Field data acquisition should be conducted to evaluate the flow and dynamic pressure.
Polymer lubricated bearings are some of the most difficult to analyze. The non-Newtonian characteristics of the bearings make the bearing analysis challenging. Any analysis must evaluate the viscosity as a function of shear rate and temperature to properly capture the fluid dynamics and heat transfer. The reaction center of pressure of the bearing loads are different than Newtonian bearing analysis. Field data acquisition is key to understanding the performance of the bearing.
KnightHawk Engineering has more than 30 years of experience analyzing polymer gear pump design and failure analysis.
For more information, visit www.knighthawk.com or call (281) 282-9200.