Boiler Feed Water Pump – Thermal Expansion and Root-Cause Analysis

You may rightfully ask what the connection is between the two topics, and I hope that by the end of the article, I will be sufficiently clear in providing a satisfactory explanation that sheds more light on the best practices for maintaining rotating equipment in high-temperature applications. But being an engineer, I kindly ask for your understanding and patience…

Let’s start with the concept of Root-Cause Analysis (RCA), which, without being exhaustive, aims to be a systematic process for identifying the primary cause of a defect encountered by your rotating equipment. There are several methodologies for RCA, and I will list a few for reference: 5-WHYs, A3 Problem Solving, Fault Tree Analysis (Apollo, Proact-Meridium, etc.), Kaizen Blitz, TRIZ, Fishbone Diagram, etc. As you can see, most methodologies known by their English names, reflecting the general lack of interest in seeking the fundamental cause of the defects we encounter. Generally, people tend to focus their attention and resources on immediately remedying the effect of a malfunction rather than the cause that led to it (for example, bearings in a rotating piece of equipment are usually replaced instead of analyzing the causes that led to their failure—mechanical looseness, lubricant contamination, electrical discharges in the bearings, etc.).

Now, everyone understands the concept of thermal expansion, where equipment changes its physical dimensions when subjected to a heating or cooling process (here, the term is thermal shrinkage). For instance, in the case of a Boiler Feed Water Pump rotor measuring 1,500 mm in length, we can have an increase in length of up to 2.5 mm when heated from 25°C to 135°C. This increase is remarkably significant, especially when we consider the very small clearances in the axial bearings and the labyrinth seals of the pump. Interesting, isn’t it?

Similarly, in the case of the casing of the same feed pump, we must consider dimensional increases for achieving precise shaft alignments. A more detailed discussion on this topic will be developed later. Rotofix Solutions has the perfect expertise for your precise shaft alignment, complemented by the EasyLaser XT440 devices at your disposal, which can work wonders for the operation of your rotating equipment.

Returning to our topic, what is the connection between thermal expansion and Root-Cause Analysis? I will tell you that we encountered an event with a Boiler Feed Water Pump where the thermal expansion of the casing was constrained (a situation commonly found in many high-temperature applications) due to the bolts at the non-drive end site (NDE) being incorrect tightened. As you know, the area of the pump casing near the coupling (drive end / DE) is fixed point (with the help of keys, fixing bolts, etc.), while the free-end can move freely (during heating/expansion) to avoid blocking the casing’s expansion (which can lead to serious defects in the axial bearing or rotor seizing within the casing). After conducting a Root-Cause Analysis (RCA), it became evident that the expansion was impeded because the bolts at the free-end were tightened without maintaining a clearance of 0.05 mm between the hold-down bolts washers and the casing (recommended by the manufacturer) when the pump casing was cold (at ambient temperature).

This pump operated acceptably for 40 years, but immediately after initial commissioning (40 years ago), the equipment experienced various defects (mainly, rotor seizing within the pump casing once heated up), forcing the maintenance team to increase clearances in the rotor and stator labyrinth seals (these clearances doubled when compared with the minimum clearance recommendations suggested by the API Standard 610 – Centrifugal Pumps for Petroleum, Petrochemical and Natural Gas Industries. As mentioned, a simple but incorrect solution (labyrinth seals clearance increment) is often more readily accepted, which only addresses the symptoms or effects (instead of directly attacking the underlying cause of the problem). It is worth noting that the solution of increasing labyrinth clearances implemented at the initial commissioning was never approved by the pump manufacturer.

Thus, the Root-Cause of the problem is the thermal expansion constraint of the pump, while the implemented solution merely addressed the effects of rotor seizing in the Boiler Feed Water Pump casing.

It would be interesting to study how many cases you have seen where only the effects of a defect are addressed instead of definitively solving the problem by addressing the root-cause!

I look forward to your comments.