Understanding Canned Motor Pumps

EP Editorial Staff | September 1, 2008


Of course, volumes have been written as to why pumps fail. Also, it is no stretch to assert that, when pump fires result from such failures, a mechanical seal is usually involved (Fig. 1).

Mechanical seals often fail as a consequence of prior bearing distress. In some instances, seal manufacturers supply products with close clearances between the periphery of rotating seal parts and the diameter (bore) of stationary seal parts. We believe that users should specify seals for hydrocarbon products to comply with API (American Petroleum Industry) Standards. We also believe that deviations from the user’s specification or applicable guidelines (such as API-682) should be brought to the purchaser’s attention.


There are, however, other ways to avoid seal failures. One of these would be to specify, whenever applicable, canned motor centrifugal pumps. Also called “hermetically sealed” pumps, they do not incorporate mechanical seals.

History of canned motor pumps A canned motor pump cross-section is shown in Fig. 2. The development of these centrifugal pumps is closely associated with the expansion of nuclear power generation technology. As of the early 1950s, safety considerations led to the development of hermetically sealed operating loops. It was then that the design principle of the canned motor— known since 1914—found practical application.

The chemical industry’s recognition of the advantages of these pumps followed soon thereafter. Indeed, the additional demand created a widespread economic base for the manufacture of canned motor pumps.

By the 1960s, canned motor pumps had evolved to the point of standardization. Expanding populations and rising standards of living in both the industrial and developing countries required innovative technologies to solve the demanding and progressively urgent problems of protecting the environment. Increasingly, canned motor pumps became part of the answer.

We readily acknowledge that, in the last three decades, much progress has been made in the field of mechanical shaft sealing. Yet, beginning with the mid-1970s, environmental considerations, industry consolidation and automation of processes have become increasingly important. It is in this context—in a wide range of fluid movement tasks—that the more traditional types of sealing either provide inadequate safety or present pollution and loss concerns that can no longer be tolerated.

In some instances, the cost of seal support and monitoring systems is disproportionate to the potential success. It is only fair to point out the existence of services that simply cannot be performed using “open” or conventionally sealed centrifugal pumps. Absolutely hermetic transport of fluids using centrifugal pumps is only possible where the torque applied to the pump rotor is generated externally. To satisfy this requirement, a rigid external stator system using either electromagnets or permanent magnets is needed. In some industries, conventionally sealed pumps can endanger human life and physical assets. Thus, traditionally sealed pumps are not always “best available technology.” Accordingly, wherever the state-of-the-art makes this protection feasible, the means employed must be aligned with the results achieved.

And so, keeping in mind the limitations of “open design” pumps (those with packing or mechanical shaft seals) that can contribute to air and water pollution, we should endeavor to be thoroughly acquainted with state-of-the-art of hermetic drive techniques for centrifugal pumps. The facts may surprise us.


Design and functional description As extremely environmentally sound machines, canned motor “hermetic” pumps are now very widely used in Europe and Japan. While making inroads in the United States, lost ground needs to be recovered in our quest for competitiveness. That said, wherever it is necessary to move dangerous, toxic, polluting, expensive, caustic, potentially explosive, high-temperature or low-temperature fluids, canned motor pumps deserve very close consideration.

The canned motor combines the well-understood hydraulics of centrifugal pumps with the equally wellproven three-phase induction motor. The hydraulic section is directly connected to the drive motor. A pipelike sleeve or “can” is inserted in the magnetically-bridged gap between rotor and stator. The “can” absolutely and hermetically separates the rotor chamber from the pressurized fluid pumping environment. In other words, the “can” is the boundary between the liquid-enveloped pump rotor and the non-wetted stator chamber (Fig. 2). The “can” thus separates the motor into two functional areas; it represents the hermetic sealing element of the pump assembly. In essence, the torque required for shaft rotation is transferred via the can, which consists of a non-magnetic material, by electromagnetic means. This type of drive does not require a shaft aperture through the fluid-containing (usually pressurized) housing; there is thus no need for dynamic gaskets or mechanical seals. The necessary static gaskets are generally problem-free but, in special cases, may be replaced by welded connections.

Canned motor pumps, therefore, are fully hermetic pump units. The pump section can be of single or multistage design. The pump impeller (or impellers on multistage pumps) is mounted at the overhung end of a shared pump-and-motor shaft. The performance parameters of these pumps now correspond to the stipulations of their main areas of application—the chemical and refining industries. At present (2008), the upper power limit is the vicinity of 600 kW. In order to approach as closely as possible the dimensional and performance-related envelopes of large numbers of standard centrifugal pumps (DIN 24256 or, respectively, ISO 2858 used in the chemical industry), many thousands of canned motor pumps in service today are available with the standard hydraulics of this pump range. The identical external dimensions of DIN and ISO hermetic and “traditional” centrifugal pumps allow rapid conversion from conventional to hermetic pumps. Needless to say, this allows reducing the spare parts inventory requirements of any modern facility.

There are, however, substantial additional advantages. These advantages, as well as the impressive efficiency and MTBF statistics of hermetically sealed API-compliant pumps, are discussed in the following Sidebar section, co-authored with George Dierssen, of IndustryUptime.


As equipment reliability consultants, we clearly can see at least five (and more likely 10) definable benefits of canned motor pumps over traditional API-610 pumps. These benefits belong to one or more categories that ultimately translate to safe, reliable, low-maintenance, low-installed-cost, as well as environmentally compatible service. To enumerate some benefits:

  • Positive secondary containment, no uncontrolled leaks to atmosphere (even with failed bearings)
  • No mechanical seal
  • No alignment (applies to installation and maintenance activities)
  • No lubrication (no oil)
  • No foundation or grout

At present, each of the five refineries in the San Francisco Bay Area has one or more canned motor pumps in highly satisfactory operation. While representing from near-zero to perhaps 2% of the pump population, we now believe canned motor pumps (CMPs) are probably applicable to 50% of the pumps at a typical refinery site. There are, of course, restrictions (i.e. thermal shock, dry running, slurries, etc.), which, in most cases, can be handled by engineering controls. Nevertheless, it seems that CMPs are the ideal choice for many HP (hydrocarbon processing) services. This is especially so since today’s plants have high expectations regarding plant uptime and operational safety. Although these expectations are reflected in a solid standard—API-685—the meager representation of CMPs in U.S. plants is puzzling.


Barring any unusual mathematics, one of the co-authors (of this Sidebar section) was surprised to discover that the average mean time between repairs (MTBR) for CMPs—factoring in every one of the many thousands made by two separate major manufacturers—is 7.5 years. This leads us to wonder as to refining industry data overall. Our understanding is that 80-90% of Japanese plants use CMPs (one manufacturer apparently ships 1700 pumps per month!) and 60-70% of European plants use either magnetic-drive pumps or CMPs. Thus, we followed up on the question. Our evaluation and substantial input from a respected CMP manufacturer can be summarized in a number of important points.

  1. A successful European manufacturer has experience with CMP power inputs of 700 hp and more. These are offered in many different configurations, with or without coolers, with or without separate bearing lubrication loops, with one or two (or more!) stages, etc. Space constraints will allow us to show only one of these (Fig. 3). Additionally, rather high fluid pressures are now commonly achieved by CMPs. Although independent, a German manufacturer uses Swiss pump technology for impeller hydraulics. Separate (clean) slipstreams lubricate the sleeve bearings typically installed in CMPs from this company.
  2. In the U.S., industry still has difficulty abandoning the bad practice of installing ill-fitting piping on fluid machines. While it was originally thought a CMP might be slightly more vulnerable than an API-610 centrifugal pump, this is no longer the case for the vast majority of CMPs. In every instance known to one expert, canned motor pumps (CMPs) were designed for significantly higher allowable nozzle loads than the equivalent API-610 pumps. The European CMP manufacturer is in a position to offer allowable nozzle loads three to four times the maximum API-610 allowable loads (remember that there are no alignment issues with CMPs). Flange bolting usually is the limiting factor here.
  3. We recall that decades ago (and in sizes approaching 1000 hp), vertical process pumps equipped with soleplates floating on the pump foundation became the norm at bestof- class facilities. Likewise, the European manufacturer recommends that CMP baseplates (actually, soleplates) not be anchored, but be allowed to float with the piping. Floating soleplates offer significant savings in piping costs. On several recent high pressure projects the piping savings equaled the cost of the pumps! Still, to this day, some purchasers insist on baseplates being furnished.
  4. Years ago, some HPI (hydrocarbon processing industry) locations had been faced with labor union concerns— whether a CMP is an electrical device or a pump was the issue at the time. The CMP expert mentioned that, while this might be a problem, his current recommendation is to return the CMP units to a highly experienced repair facility in Louisiana for any necessary repairs. We believe this approach makes sense in view of the limited availability of qualified field maintenance and machine shops personnel in many plants.
  5. We also made the observation that all too often engineers listen only to the marketers of traditional pumps—“I can’t make a mistake if I just do exactly what my former boss did. He always played it safe and now he’s VP of Engineering!” The national sales manager for a prominent canned motor pump manufacturer agreed and said there is still much real reluctance to try what is perceived as “new technology”—although, of course, CMPs are considered to be a mature technology.
  6. Perhaps, and often quite wrongly, it is assumed that CMPs consume more energy than traditional pumps. A European CMP expert refuted this incorrect belief. He noted that, in over 30 years of applying CMPs, he had never seen one of these units consume more energy than the pump it replaced! While he had once assumed this was due to worn pumps being replaced or better selections (curve fits) being found, he now believes that the original units were overly optimistic on their published effi- ciency and never included many losses—such as those attributable to couplings and seals.
  7. A pump specialist related his experience with one manufacturer of conventional pumps that always tested its units with lip seals instead of braided packing or mechanical seals. Some years ago, this expert was made aware of a company that let it be known they would evaluate quotes (bids received) based on input kW. As a result, the “guaranteed” horsepower went up almost 10% over some bidders’ published curves. The expert advises users’ specifying authorities to request an input kW value for the equipment offered. He noted that prominent CPM manufacturers would be pleased to comply since they, of course, supply both pump and motor.
  8. The average user/purchaser is disappointed that he cannot send trash through bearings and close clearance components. Of course, the vendor must really educate the user on the limitations of the equipment and ask repeatedly about process stream and fluid properties. Many failures that in the past were blamed on “solids” or “dry running” were actually initiated by internal flashing of the fluid. That may have been due to lack of a good heat balance program on the equipment or just plain ignorance of the supplier. Even more important, there often has been insufficient failure analysis or follow-up by the manufacturer to determine the true root cause of a failure. One interesting factor is maximum heat rise; it occurs on a CMP after shutdown due to the latent heat in the motor. This concern always can be addressed by proper motor sizing, purging the motor after shutdown or other methods—provided the buyer and a knowledgeable manufacturer cooperate. What is a concern for some is a non-issue for others.


  1. Consider it the manufacturer’s responsibility to ask many questions and for both user and manufacturer to provide solid answers. Pursue questions on maximum (short-term) allowable temperature rise of the pumped fluid. The answers to such questions are easy to obtain. If an intelligent user knows what he’s pumping and accurately describes his process fluids, the fact that he’s about to purchase a CMP is of secondary importance.
  2. Decades ago, the superiority of CMPs was described in Kenneth Fischer’s joint Hoechst-Celanese presentation at one of Texas A&M University’s International Symposia. The proceedings of these gatherings are readily available and the earlier editions addressed many user concerns and provided some failure statistics, etc.
  3. All too often, the buyer leaves decisions to the design contractor and then encourages these firms to buy largely on the basis of cost and schedule. The contractor then buys a conventional pump from the lowest bidder. A reliabilityfocused user must step in and take a measure of responsibility for the guidance and direction needed by design contractors and purely purchasing-oriented personnel. Solid life-cycle cost studies are better than preconceived or outdated opinions.
  4. Also, and sadly, CMPs might have gotten a bad name when certain sellers oversold their merits in times past. While the merits of canned motor pumps are indisputable, there is nothing that cannot be mislabeled, misunderstood, maligned or destroyed. There are no exceptions to this rule.

As always, we invite your comments.

Contributing Editor Heinz Bloch is the author of 17 comprehensive textbooks and over 340 other publications on machinery reliability and lubrication. He can be contacted at:

George Dierssen is a principal with IndustryUptime, a firm headquartered in Benicia, CA. Among its many services, IndustryUptime helps its clients improve pump reliability, increase process availability, meet tough new emissions standards for pumps and reduce energy costs associated with pumping systems. It also provides rotating equipment application support to Dupont Corporation for Vespel® CR-6100, a non-metallic wear material that improves pump reliability and efficiency. For more information, e-mail:






View Comments

Sign up for insights, trends, & developments in
  • Machinery Solutions
  • Maintenance & Reliability Solutions
  • Energy Efficiency
Return to top