[webinar] Embracing Digital Transformation in Maintenance & Plant Operations | March 13 at 10AM EST – Register Now
Suction Bell Upgrades for Vertical Turbine Pumps
Suction Bell Upgrades for Vertical Turbine Pumps
Dr. Lev Nelik, P.E., APICS
How long should a vertical pump’s suction bell last? That was a question a plant manager asked me during a recent consulting assignment for a power plant. A pump manufacturer’s typical response to such a question is, “it depends.”
Basically, it depends upon the liquid being pumped, the operating point on a curve, the accumulated hours, the materials of construction, and so on. But is there an average value, or does a pump life vary within a very broad range? Is 30 years of life considered a very long time? Is two years unacceptably low?
If a large vertical cooling water or recirculating pump supplies water to the plant, 15 (or more) years of life should be achievable. Anything under five years is too short for most applications. If water comes in from the ocean or a gulf, the pump life (i.e. impeller, casing cone, suction bell) will be reduced, especially if a pump operates off-design, where suction recirculation is particularly damaging and the seawater is known to be very corrosive.
Often, damage to the impeller casing (cone — at immediate proximity to the vanes) is much more severe as compared to the damage of the suction bell (below the impeller) itself. In recognition of this fact, some designs are deliberately configured into a two-piece construction where the impeller casing (cone) is bolted to the suction bell, thus making it simpler to replace. For large vertical turbine pumps, these parts are expensive, and separating a shorter-expectancy-life part from a longer-expectancy-life part is a good idea.
Unfortunately, many designs configure the impeller casing and suction bell into one continuous unit because it is obviously cheaper to produce a single-piece casting. The problem here is that when the casing gets damaged, the entire part must be scrapped and replaced — and the end user pays for that.
When examining such a worn-out casing cone/ suction bell part, an effective upgrade strategy involves modifying the design by separating the part into two pieces and then bolting them back together. An example of such an upgrade was recently done by our repair shop.
Suction bells are usually constructed from iron, which is significantly less expensive than stainless steel but lacks its resistive properties. Upgrading an entire piece is another possibility, but it is often too expensive and really unnecessary. Instead, a two-piece construction will have a new stainless cone, bolted to the old bell and fitted with a stainless flange.
The result is a much better design, with stainless steel providing many more years of operation at enhanced resistance to corrosion and cavitation damage.
And should the lower portion (suction bell) eventually need repair, it can be done, saving time and money by reusing the undamaged stainless upper part.
We cannot change the single-piece past. But we can certainly improve the life of individual components, once the deficiencies of past practices are realized, understood — and corrected.
A pump’s efficiency can degrade as much as 10% to 25% before it is replaced, according to a study of industrial facilities commissioned by the U.S. Department of Energy (DOE), and efficiencies of 50% to 60% or lower are quite common. However, because these inefficiencies are not readily apparent, opportunities to save energy by repairing or replacing components and optimizing systems are often overlooked.
A pump’s efficiency can degrade as much as 10% to 25% before it is replaced, according to a study of industrial facilities commissioned by the U.S. Department of Energy (DOE), and efficiencies of 50% to 60% or lower are quite common. However, because these inefficiencies are not readily apparent, opportunities to save energy by repairing or replacing components and optimizing systems are often overlooked.
First of all, despite the fact that the PLC was designed as a direct replacement for relays, its logic is actually quite different. Relays are 100% parallel logic. Every single part of a relay control system operates simultaneously. If you were to draw several rungs on a relay diagram and put one coil on each line with no contacts on any of the lines, every relay would energize at the same time when power was applied. This makes relay logic blindingly fast by nature (its only the relay's mechanical limitations that make it slow) but it’s often a source of trouble
First of all, despite the fact that the PLC was designed as a direct replacement for relays, its logic is actually quite different. Relays are 100% parallel logic. Every single part of a relay control system operates simultaneously. If you were to draw several rungs on a relay diagram and put one coil on each line with no contacts on any of the lines, every relay would energize at the same time when power was applied. This makes relay logic blindingly fast by nature (its only the relay's mechanical limitations that make it slow) but it’s often a source of trouble
It turned out that the photoelectric sensor could only be adjusted to stay on all the time or off all the time when the cans were going by at any higher rate than Jog speed on the Seamer. We adjusted and adjusted the sensors. The sensors were replaced with new sensors. It did no good. The lines were stopped while we traced all the control and power wiring so we could try to determine if there was a problem in the wiring.
It turned out that the photoelectric sensor could only be adjusted to stay on all the time or off all the time when the cans were going by at any higher rate than Jog speed on the Seamer. We adjusted and adjusted the sensors. The sensors were replaced with new sensors. It did no good. The lines were stopped while we traced all the control and power wiring so we could try to determine if there was a problem in the wiring.
Is there is reason to suppose a properly repaired motor is less reliable than a new machine? This article will define reliability and repairability to help you answer this question.
Is there is reason to suppose a properly repaired motor is less reliable than a new machine? This article will define reliability and repairability to help you answer this question.
Ask for a modern centrifugal pump recommendation from your favorite supplier and chances are he will recommend one of the standard pump designs that conform to either the A.N.S.I., I.S.O. or D.I.N. specifications. On the surface that might seem like a good recommendation, but the fact is that all of these designs will cause you maintenance problems. Refer to the illustration prior to diving into the details of the obvious problems found within these designs.
Ask for a modern centrifugal pump recommendation from your favorite supplier and chances are he will recommend one of the standard pump designs that conform to either the A.N.S.I., I.S.O. or D.I.N. specifications. On the surface that might seem like a good recommendation, but the fact is that all of these designs will cause you maintenance problems. Refer to the illustration prior to diving into the details of the obvious problems found within these designs.
Every component of a bolted flange joint has a maximum allowable stress level. The mating flanges will begin to rotate or warp at a defined stress threshold. The studs or bolts of a given specification also will yield or be stressed past their elastic properties at a defined level. Non-metallic and semi-metallic gaskets will crush under excessive applied stress loads.
Every component of a bolted flange joint has a maximum allowable stress level. The mating flanges will begin to rotate or warp at a defined stress threshold. The studs or bolts of a given specification also will yield or be stressed past their elastic properties at a defined level. Non-metallic and semi-metallic gaskets will crush under excessive applied stress loads.
