...water pump experts. Why do the mechanical seals on pumps purchased elsewhere frequently leak, whereas the seals on pumps manufactured by Shanghai Shenyin Pump Co., Ltd. rarely do? Upon dismantling the other pumps, it was discovered that the mechanical seals showed no signs of wear. So, what could be causing this phenomenon? Today, our pump experts at Shanghai Shenyin Pump will provide a brief analysis of this issue.
Case Study: Mechanical Seal Leakage
Consider a newly installed pump handling a black, viscous oil containing a large quantity of solid particles ranging from 2 mm to 10 microns in diameter. The operating temperature is 210°C (with a viscosity of 50 cP at this temperature), and the fluid is pH-neutral, causing no corrosion to steel or common plastics. Operating parameters are as follows: inlet pressure 0.1 MPa, outlet pressure 0.6 MPa, flow rate 945 m³/h, shaft power 147 kW, and rotational speed 1480 rpm.
The pump utilizes a tandem double-end mechanical seal (assembly and component details: 1. shaft sleeve spiral-wound gasket; 2. shaft sleeve; 3. metal bellows mechanical seal; 4. gland seal ring; 5. inner gland; 6. connecting screw; 7. outer gland; 8. cylindrical pin; 9. positioning block; 10. bolt). The flushing plan is PLAN 32 + 52; the external flush fluid is clean wash oil at over 100°C, and the barrier/buffer fluid is No. 46 turbine oil (yellow in color). After three months of operation, significant black oil leakage was observed at the mechanical seal; however, the barrier fluid remained yellow, and leakage ceased when the pump was stopped. Upon inspection and disassembly, the sealing rings at both ends were found to be intact, the springs were not binding, all O-rings and gaskets protruded correctly from the seal chamber, and the pump casing showed no porosity defects. So, where is the leakage coming from, and what is the cause? We invite you to consider the problem and ask questions; the answer will be revealed later.
Two main hypotheses have been identified, though each is contradicted by counter-evidence:
1. Leakage occurring via the shaft sleeve area. However, since this involves a static seal point, the leak *should* occur regardless of whether the pump is running or stopped; yet, the leakage only happens when the pump is running and stops when it is shut down.
2. The medium is leaking externally past the two seal rings. One would expect a leak at the first seal ring to immediately contaminate the seal oil, yet the quality of the seal oil has remained unchanged.
What is even more baffling is that a visual inspection of all the parts removed during maintenance revealed no obvious defects. Perhaps my perspective is limited, but in my years of experience, parts that fail after a period of operation usually show visible signs of damage. If the parts themselves are not the problem, then where does the issue lie?
In fact, the root cause was not fully understood during the initial maintenance; however, to meet production demands, all seal-related components were replaced before the pump was reassembled. Yet, after just one month of operation, the mechanical seal began leaking again, exhibiting the exact same symptoms: leaking only when running and stopping when the pump was shut down, with no obvious defects found on the removed parts.
Approach to resolving the mechanical seal leak
From the initial observation that the seal oil remained uncontaminated during the leak, it could be inferred that the leakage originated at the shaft sleeve. An analysis of the shaft sleeve seal structure (see diagram below) reveals no inherent design flaws; indeed, this design offers advantages over the alternative of installing an O-ring at the inner end of the sleeve. First, it allows for the use of non-rubber sealing elements—such as copper gaskets or spiral-wound gaskets—which naturally have a longer service life than O-rings, as the latter are prone to aging. Second, it eliminates the risk of scratching the O-ring during installation; with the inner-bore O-ring design, the O-ring is easily damaged by impurities or uneven force when the sleeve is slid onto the shaft. In fact, many pumps operate successfully using this type of seal.
The unique factor here is the pump's large impeller. Although the head is relatively low, the combination of a low rotational speed (1480 rpm) and a high flow rate (945 m³/h) necessitates a large impeller—approximately 550 mm in diameter and 80 mm thick. As shown in the diagram, the large impeller generates significant axial pulling force during operation, causing the preload on the gasket at position "1" to decrease; the preload is restored when the pump stops. Since the gasket at position "1" is a spiral-wound type, it lacks elastic recovery. Repeated start-stop cycles and process fluctuations cause the gasket to be alternately compressed and released, eventually leading to failure. This results in an excessive gap during operation while maintaining a good seal when the pump is idle—explaining why leakage occurs only when the pump is running.
Having identified the root cause, the appropriate remedy is to switch to an O-ring installation at the inner end of the shaft sleeve. API 682 recommends placing the O-ring as close as possible to the locating shaft shoulder to minimize friction and compression damage during installation. However, the original sleeve design lacked this O-ring feature and was consequently too thin to accommodate a machined groove for one. Therefore, a new shaft sleeve must be manufactured. This new sleeve will include an extended section near the mechanical seal locking clip to house the inner O-ring. The only critical requirement during installation is to exercise caution to prevent the shaft from scratching the O-ring. Since the new mechanical seal had not yet been manufactured—only the shaft sleeve was ready—all other components used during the second overhaul were reused parts. After reassembly, the unit operated flawlessly for the year and a half I remained in that position. We hope this guide on resolving mechanical seal leaks proves helpful to you.