By Larry Bachus, The Pump Guy

Industrial Safety and Industrial Reliability were born together in the early 1970s, as twin children of Aviation Safety and Reliability. The aviation industry wrote the book on safety and reliability as the airline companies converted their fleets of passenger planes from propeller engines to jet engines.

Aviation safety revolves around passenger safety. Aviation reliability revolves round equipment reliability. However, the goal of passenger safety is also the goal of equipment reliability, that is: a better experience.

Anything — night time runway lighting for example — that improves flight safety also improves flight reliability. Scheduled jet engine preventive maintenance improves both equipment reliability and passenger safety. A passenger jet that stays in the air and arrives at its destination on time without incident or accident is both safe and reliable. One hand washes the other hand.

The distinguishing factor between the aviation safety and reliability programs, and the industrial safety and reliability programs is: The workers (the flight crew) and customers (the paying passengers) ride inside the flying equipment. If the jet is reliable, the crew and passengers are safe.

In industry, with very few exceptions, the workers don’t ride inside the running equipment. Exception: The crane operator rides inside the crane. The bus driver with passengers ride inside the bus.

In industry, worker safety and equipment reliability have different goals. They are separate departments with separate goals.

In 2018, we can say that most process plants are safe. Companies brag about their safety statistics. The company officers enthusiastically support the safety program. The safety engineer proudly publishes the safety statistics in a company newsletter or on a prominent billboard. Everyone (visitors, contractors, part-timers, secretaries, company officers) gets training on “safety”. Everyone is encouraged to report an unsafe action or situation. Everyone practices and contributes to safety.

Notice the trough under the seal chamber to carry-away leakage. The astute seal salesman calls this “Proactive”.

However, too many process pumps are not reliable. Too many reliability programs are stalled. Often, company officers don’t really support the reliability program. The reliability engineer doesn’t publish reliability statistics on pumps and other rotating equipment. Who, besides the reliability team, goes to the reliability meetings? No one is encouraged to report an unreliable action or situation. No one knows how to contribute to reliability.

Most process plants divide the assets into rotating equipment (pumps, motors, fans, compressors, gear boxes, turbines, air extractors, etc) and stationary equipment (tanks, pipes, valves, heat exchangers, etc). Process pumps drag-down the reliability statistics.

Most reliability people know and use the term ‘Bad Actor Pump’. Where are the ‘Bad Actor Fans’? Where are the ‘Bad Actor Transmissions’? You see, fans, motors and gearboxes don’t drag down the reliability statistics. Process pumps with mechanical seals drag down the reliability statistics.

The mechanical seal was patented 115 years ago in 1903. 1903 was the same year the Ford “Model T” automobile, and the Harley Davidson motorcycle went into production. The Wright Brothers initiated powered flight in 1903. And the screw-in electric light bulb was patented in 1903.

115 years later, electric lighting is mostly perfected even as it evolves into LEDs. Commercial aviation is mastered, safe and reliable. But over a century later, too many process pumps are not reliable because too many mechanical seals fail mysteriously, prematurely.

The automobile companies make large and small vehicles (cars, trucks, convertibles, off-road vehicles, minivans, motorhomes, sports cars, etc) for the needs of the driving public. You’d think the mechanical seal industry would make reliable seals for the needs of the process industry.

How about a mechanical seal for pumps in cavitation? We need a mechanical seal for deadheaded pumps and dry-running pumps. We need seals designed for pumps that operate for extended periods on the extreme left or right of the performance curve. We need a seal for starved pumps. We need mechanical seals designed to handle production upsets. These are the real reasons that so many pumps suffer premature seal failure. Why doesn’t the seal industry recognise this?

And yet, the seal manufacturers continue to promote ‘easy installation’, ‘interchangeable spare parts’, rebuild kits and ‘off-the-shelf availability’. How about longer service life in a process pump? We don’t really need another seal that can be installed by an apprentice technician in five minutes… every four months.

“Service” is the word used at a stud farm for racehorses. The breed stallions “service” the receptive mares. Maintenance engineers are tired of being “serviced” by their seal suppliers.

Generally, pumps don’t break. Pumps leak! The impeller doesn’t split into five pieces. The shaft sleeve doesn’t fracture into three pieces. The seal chamber doesn’t crack. The volute casing rarely splits down the middle.

A process pump can overheat and continue pumping. Some pumps suffer cavitation for months and continue pumping. Pumps can vibrate wildly and continue running. This is called “Running-to-Failure” (a concept that deserves its own article.).

Pumps leak because the seal fails prematurely. Admittedly there are other pump problems (mentioned in the previous paragraph). Eventually, the excess heat, incorrect piping, wild vibrations, mechanical loading, cavitation, misaligned bearings, dry-running, deadheading and operational abuse will stress the seal until it unexpectedly leaks and fails. Then the pump goes into the shop.

Traffic jam at the seal rebuild shop. These cartridge seals suffered the same failure from the same pumps. Many process plants don’t analyse failed seals.

Mechanical seals leak either at the seal faces, or through the secondary elastomer component, normally an O-ring. I promise you the process liquid is not extruding through the seal’s stainless steel body, or through the set screws, springs or drive pins. The drip doesn’t find a microscopic pathway through the non-porous ceramic ring. All leakage is through the union at the seal faces, or through the O-rings and/or gaskets.

40 years ago, no seal manufacturer was intimately associated with a pump manufacturer. But now, with corporate mergers and “team-playing” mega-conglomerates, a few companies have a pump manufacturer and a seal company in the corporate portfolio. The seal manufacturers and pump companies must jointly direct their efforts toward the seal-face union, and the secondary elastomers.

Evidence suggests the seal manufacturers aren’t paying attention. We haven’t seen any great leaps in mechanical seal technology in recent years. There is more action in eliminating mechanical seals. Here are three examples.

Elastomer technology

Perfluoroelastomer O-rings have been around for almost 40 years. This elastomer compound can resist a broad range of aggressive chemicals and withstand temperatures up to about 300°C without degradation. To date I’m not aware of a successor O-ring. Where is the 400°C O-ring? Or the 600°C O-ring? Where is the O-ring with even broader chemical resistance?

Sealless pumps

This is actually a misnomer. Mag drive and canned motor pumps have seals, (O-rings and gaskets) but they don’t have a mechanical seal on the rotating shaft. We should correctly call them “mechanical sealless pumps”. Mechanical sealless pumps are available in a wide range of designs, covering PD and centrifugal types. Mechanical sealless pumps are thought to be a recent development. Actually, they are the oldest of designs. The Archimedes Screw had no seal on the rotating shaft.

This failed mechanical seal will send this hot water pump to the shop.

Dry gas seals

The first dry gas mechanical seal is close to 70 years old (patented in 1951). These dry gas seals have performed well in rotary compressors since then. There are a few reasons for the success in compressors.

The range of gases that are typically compressed is limited. Examples are air, nitrogen, oxygen, hydrogen, hydrocarbon gases (propane, butane, natural gas, acetylene, etc), refrigeration gas, and CO2 as fire suppression gas. Also, the gas is normally filtered and free of solid particles and other contaminants as it passes through the compressor.

Containing a process gas with a compatible, pressurised barrier gas is relatively easy. Containing a dirty, abrasive, aggressive, non-compressive liquid with a barrier gas is another animal.

In the late 1960s, someone said, “Because dry gas seals work so well in compressors, why don’t we put dry gas seals into centrifugal process pumps?” To be politically correct, this seal has a “Long and Slow (30+ years and still counting) Learning Curve” in process pumps. There, I said it. Well someone needed to say it.

Dry Gas seals have an uphill battle in pumps as long as:

  • Control room operators can’t interpret a pump performance curve
  • Process pumps lack adequate instrumentation
  • Process pumps are allowed to operate indiscriminately all over their performance curves in dynamic process systems

It makes no sense to install a mechanical seal where the axial gap required at the seal faces is stricter than the permissible axial play in the shaft bearings. What’s the purpose of a mechanical seal with a sophisticated support system that costs more than the seal itself? Isn’t this like a tyre that costs more than the car?

The fuse is the weak link in an electrical circuit. A burned fuse is the physical manifestation of a problem in the electrical circuit. The electrician doesn’t change the same burned fuse 20 times in the same circuit. The electrician doesn’t blame the fuse for failing. After two or three burned fuses, the electrician knows the problem and the solution is in the electrical circuit and not in the fuse or fuse box itself.

Likewise, the mechanical seal is the weak link in a hydraulic circuit. Numerous failed seals in the same pump is the physical manifestation of a problem in the hydraulic circuit. So, why does the maintenance engineer change the same failed seal 20 times in the same pump? Why does the maintenance manager blame the failed seal? The reliability engineer should do a failure analysis on every seal that fails prematurely to be sure the seal was specified properly and installed properly.

After two or three premature seal failures in the same pump, the reliability engineer should search for the problem and the solution in the hydraulic circuit.

The Pump Guy is Larry Bachus, a maintenance practitioner, inventor, lecturer and contract mentor based in Nashville, Tennessee. Mechanical seals, seal failure and failure analysis are discussed in Larry’s book Everything You Need to Know about Pumps. Larry is a retired member of ASME, and speaks fluent English and Spanish. You can contact Larry at

Find Seals Related Companies In The Pump Industry Capability Guide

Related articles

©2024 Pump Industry. All rights reserved


We're not around right now. But you can send us an email and we'll get back to you, asap.


Log in with your credentials


Forgot your details?

Create Account