Pump system failure

From Ocean Navigator #58
January/February 1994
Pumps on board the average boat are usually misunderstood, mistreated and subjected to unreasonable demands. Perhaps because of this, pumps stop pumping with frustrating regularity. Often the pump is wrongly blamed for failing when there were other factors at play that caused it to fail.

In order to correctly diagnose any pump failure, a thorough understanding of how a pump works is needed. Additionally, it is helpful to know how a pump’s performance is likely to be affected by its environment.

Pumps are mechanical devices that employ simple principles of physics to move fluids in directions they would not normally want to flow by themselves. The physical constant which allows us to do this is atmospheric pressure. All pumps (those at least encountered in normal use in terrestrial applications) operate by displacing fluid and creating a vacant space. This space is then refilled by atmospheric pressure acting on the surface of the reservoir or source, and emptied again by the continued action of the pump.

Average atmospheric pressure at sea level is 14.7 pounds per square inch (psi) which is equivalent to the pressure exerted by a column of fresh water 34 feet in height. The theoretical maximum height, then, to which fresh water can be drawn by a pump is 34 feet. The practical limit is much lower, about 25 feet, depending on the type of pump and its construction. The height to which water can be pushed by a pump is limited only by the type of pump, its construction, and its power.

Although they all depend on atmospheric pressure to function, there are three distinct mechanical principles which separate pump types.

Reciprocating pumps: This is the oldest pump known to man. In its rudimentary form, the lift pump, it consists of a piston which is drawn through a tube to create a vacuum which is then filled by atmospheric pressure on the source – a well for example. The cylinder is fitted with two valves, one which opens on the charging stroke to admit fluid and one which opens on the discharge stroke to allow the fluid to be sent to its destination.

The action of the reciprocating pump makes it a positive displacement pump, i.e., it moves a defined amount of fluid at each stroke. It is also self-priming, because as long as it is operating within the limit imposed by atmospheric pressure, fluid will always flow into the evacuated space.

The reciprocating pump has evolved into many forms, many of which employ diaphragms instead of a piston, but still operating under the same principle of alternately filling and evacuating a space with a reciprocating motion, and employing valves which alternately open and close.

Rotary pumps: Various versions of the rotary pump employ a rotating impeller, or gears, to create a space, entrain fluid and expel it. As with the reciprocating pump, the creation of a finite space which is filled by a specific amount of fluid makes this both a positive displacement type of pump and a self-priming one.

Types of rotary pumps commonly found in marine applications are the flexible impeller (e.g., engine raw water pump) and the vane pump (fuel transfer pump). Gear pumps are used in hydraulic systems which require high discharge pressures.

Centrifugal pumps: The centrifugal pump is not a positive displacement pump and is not self-priming. Fluid is admitted at the axis of a volute-shaped housing where an impeller rotates the entrained fluid, imparting velocity (and kinetic energy). The fluid is then carried to the outlet by centrifugal force, creating a vacuum at the axis that draws in more fluid.

Unless primed, the impeller will be spinning only air, which will not achieve sufficient momentum to overcome static lift and draw in a heavier fluid. A centrifugal pump, therefore, must always be installed below its source unless fitted with a check valve or other device which prevents it and its supply piping from losing prime.

There is a wide variety of pumps available on the market, which indicates that the evolutionary process has caused specific equipment to have been developed for specific uses. Any one pump type will not serve all tasks effectively.

Since such a great choice exists, it places a burden on an engineer responsible for specifying a pump for any given application to become knowledgeable about the characteristics required to fulfill that specific need. If he gets it right, the pump will serve loyally for as long as it receives the necessary maintenance.

But many pumps do not serve loyally. Many pumps fail repeatedly when, on the face of it, they have the appropriate characteristics for performing the job adequately. Why? Since there is no short answer, we’ll have to be content with a long one.

The long answer lies in the study of hydraulics, a very complex and sometimes inexact science. Certain areas of applied hydraulics, for example, the control systems for passenger airliners, are the result of heavily funded research, often with government backing. The area known as “boat plumbing” is under-researched, under-funded, and hardly understood. It is also, for the most part, under the floorboards and out of sight, therefore receives precious little attention until it fails.

Elementary hydraulics, though, can be understood and applied even in the simplest of plumbing systems, to the benefit of all concerned.

All pumps used on board a boat comprise two parts, the pump itself and the pump driver. The driver is usually an electric motor of some kind, though in the case of a manual bilge pump, the driver is the person on the end of the handle.Driver failures

Failure of a pump assembly can arise from a breakdown of either pump or driver. More often than not, it is failure of the driver that draws attention to a problem for which the pump takes the blame. However, the failure of the driver in itself is frequently symptomatic of a problem in the whole system, which includes the plumbing. If a pump or its driver stops working, it is a signal that the whole system in which it serves needs to be analyzed for factors that act unfavorably on the pump.

When a pump ceases to function, the simple solution is to replace it. If it happens again, too soon to be the result of reasonable wear and tear, the pump should not be discarded again, but taken apart and examined. At the same time, the entire plumbing run and its accessories needs to be examined.

The first step is to disconnect the plumbing (after first securing the rest of the system—closing any seacocks, etc.) and try running the pump, briefly. If it appears to run well, the problem probably lies in the plumbing, which should therefore be checked for blockages, clogged strainers, kinked piping, etc. (Be sure to check that the pump will actually pump water by removing it from its location and connecting it to an electricity supply with the intake connected by hose to a bucket of water.)

If the failure is in the pump assembly, the next step is to disassemble pump from driver and determine in which component the fault lies. Check to see if either one has seized, for example, or is hard to turn. If the motor is seized, it could be due to stalling and burning, or poor maintenance, or salt water damage.

The type of electric motor used to drive common pumps will give up the ghost for one of three primary reasons: the brushes are worn out, often a sign that the motor has been operating at an unduly high current; the motor stalls because of too high an effort required to operate the pump; the motor suffers stator lock due to low supply voltage.

The manufacturer or supplier of the pump should be able to furnish information on expected brush life which, when compared to the record of use of the pump in question should answer whether or not it has fallen short of its life expectancy. If it has died young, it is an indication that it has been working too hard, using more than the desired current to maintain its operating speed.

In the case of a rotary pump, this usually means that the manufacturer has rated the pump as delivering at a certain rate, measured in gallons per hour (GPH). The motor was chosen to operate the pump at a speed that will provide the desired GPH. The motor wants to turn at that speed and will attempt to do so, overcoming whatever resistance it meets and drawing whatever amperage that effort takes. If that current is higher than the design current, the motor will run hotter and brush wear will accelerate.

The cause of high resistance is either the pump being of too large a capacity for the service, or the service offering too high a resistance to the pump. This may sound contradictory, but on close inspection it is quite logical.

Take, for example, a raw water cooling supply for a refrigeration unit. This unit ran quite happily for a few seasons in New England, but when the owner decided to take the boat to the Caribbean the higher ambient seawater temperature created a falling off in efficiency. The owner replaced the pump (which happened to be of the rotary type) with one of higher capacity, to increase the flow through the heat exchanger. However, the pump didn’t last the season.

It turns out that the increase in flow rate incurred a corresponding increase in frictional resistance, so the pump had to work against too high a discharge head. A case of a pump too big for a system. The plumbing size should have been increased along with the pump capacity.

On the same vessel, a centrifugal pump (on paper the right specification) supplied cooling water to three air conditioning units. The two aft ones, which happened to be close to the intake, worked satisfactorily, but the forward unit never quite got down to temperature. When the supply valves were balanced so that the forward unit functioned satisfactorily, the aft ones lost out.

The problem was that the pump was nominally correct, but the plumbing it had to drive water through was too long and indirect, especially running to the forward evaporator which was mounted high in the boat. A rotary or reciprocating pump trying to supply the system would probably have run hot and eventually died. The centrifugal pump just kept on running without delivering the necessary volume, because the output of such a pump is affected by the discharge head.

The solution was to re-tune the plumbing, increasing the diameter, especially on the run forward, to reduce resistance. (See section below on hydraulic calculations and how to analyze a plumbing run to determine where a problem lies.)

The motor will stall (or suffer from locked stator) when either the opposing load is so high as to prevent the motor from turning or the supply voltage is too low. In either case, the motor sits there humming, drawing current, and heating up until it blows a fuse or the battery runs down. It can suffer permanent damage from overheating, even without drawing enough current to blow a fuse, because, obviously, its built-in cooling fan is not able to do its job. (A hot environment, such as an engine room, can affect the life of a motor and accelerate failure, particularly if it is already running on the hot side due to high load.)Electrical problems

In the event of a motor lock-up, especially when the pump has not previously shown signs of giving trouble, it is important to check the wiring. A loose or corroded connection can cause sufficient voltage drop to prevent a motor from starting up. Other culprits are weak battery voltage and wiring of inadequate gauge.

If there is no fault in the electrical supply, then the problem is in the pump or its associated plumbing. If the pump is seized it could be due to dry bearings, the pump running dry, an impeller damaged by entrained debris, or perhaps too much back pressure.

Raw water pumps are sometimes subject to damage from sand and silt, especially when operating in shallow, or estuary waters. A pump that is working against too much resistance can receive damage to its impeller, bearings and wear surfaces, which will eventually result in a dropping off of performance. Many pumps need the entrained water for lubrication; and if they run dry, they will rapidly go down the path to self-destruction.

Also, a pump which loses its prime, especially if it is of the rubber impeller type, will suffer wear every time it is started up. This problem can be eased if the piping is arranged so that the pump is at a low point and so remains wet when at rest.

These conditions occur in rotary pumps, but reciprocating, or diaphragm, pumps can have an additional set of symptoms. For example, the pump might run on forever without delivering anything. The reason for this might be as simple as a disconnected drive, but it could also be as a result of damaged diaphragms or choker valves held open by debris.

Diaphragm pumps can be subject to damage simply because of the way in which they operate. Consider a single chamber pump. On each discharge stroke a slug of water is sent down the pipe. On the charging stroke nothing follows that slug so it comes to a sudden halt. The water, therefore, moves in a series of pulses which send corresponding shocks into the pump and the piping. These shocks have the potential to cause damage.

Multiple chamber pumps generate a smoother flow, but this consists simply of more but smaller impulses. To prevent damage, the pump must be securely mounted on flexible feet, and the attached plumbing must impart no load at all on the pump connections. A loop of flexible hose will reduce the vibration transmitted to the plumbing.

A centrifugal pump is less prone to the same kind of damage as positive displacement types. The delivery rate is dependent on the head it is working against. While running at the same speed, the flow rate will diminish as the head increases, even to the point that it will continue to run with the discharge closed off. Loss of performance, though, will result from air leaks into the pump caused by a worn bearing or gland.

In general, pumps need to be used. If left idle for long periods, gaskets can dry out, causing air leaks, rubber components can stick to their housings, causing damage on start-up, and corrosion can build up. Regular use keeps these problems at bay.

Extreme temperature changes will set up stresses where differing materials join, for example where metal hose barbs are bonded into a plastic pump body, leading to vibration damage, especially in diaphragm pumps.

Accessory equipment such as float switches can malfunction, causing a pump to die by drowning or by running dry.

It is difficult to tell whether pump damage is spontaneous or the result of problems endemic in the associated plumbing, but not readily apparent in a visual inspection. If everything looks as though it should work, but we are still left scratching our head, an initiation into elementary hydraulics is called for.

The physics and terminology involved in the calculations for pumping systems are complex and obtuse. However, we are concerned mainly with simple systems which pump water, so the following information will serve for most normal purposes in recreational boats and small commercial vessels.Plumbing and friction head

All plumbing offers resistance to flow, and a quick glance at the accompanying tables will show that smaller diameter pipes offer a greater resistance, and that resistance does not increase linearly with increase in flow rate.

Diameter and velocity play the greatest parts in this area of hydraulics. Resistance is proportional to the square of the velocity, and velocity is a function of the cross sectional area of the pipe. This makes resistance a function of pipe diameter to the fourth power.

For a given flow rate, therefore, measured in gallons per minute, a half-inch pipe will theoretically offer 16 times the resistance of a one-inch pipe. Other factors do come into play to affect this relationship, but it is immediately obvious that a larger diameter pipe is usually better than a small one.

There are practical limits to the sizes of pipe we can fit in a boat, since we are constrained by space and budget. It makes sense, though, when we encounter a recalcitrant problem in a system with a pump, that we take a look at the plumbing and see if the cause lies there.

Straight pipe offers some level of resistance, and the additional resistance due to elbows, valves, strainers, and other interruptions is usually measured in terms of equivalent feet of pipe. Thus, a standard elbow in 3/4-inch pipe is equivalent to six feet of straight pipe. It is a fairly simple matter, therefore, to follow a plumbing run, measuring lengths, and counting elbows, etc. to figure out its equivalent length of pipe. Reference to the pump manual will turn up the flow rate of the pump in question and a calculation will establish whether pump and plumbing are properly matched.

For example, a submersible bilge pump doesn’t seem to meet the dewatering rate promised in the literature. This is not a straightforward calculation, despite there being nothing on the suction side to consider as the pump is immersed when operating. The delivery rate of a centrifugal pump is dependent on the resistance (head), but the factor of resistance due to friction head is dependent on the flow rate. This is, therefore, a trial-and-error calculation. We can, at least, measure the static head (the vertical distance from the pump outlet to the highest point in the plumbing) and the friction head.

The pump in question is mounted at the bottom of the sump, three feet below the waterline, and discharges via a loop (common practice on a sailing vessel) three feet above the waterline.

The discharge is through 1 1/8-inch piping and runs to the transom, a horizontal distance of 20 feet. However, by the time the pipe has made its way around the furniture and other fixtures its total length is 35 feet. There are a total of five elbows, including the loop and an elbow fitted at the discharge.

The head vs. flow chart shows that the pump will deliver 1,200 gallons per hour (gph) at 0 feet of head or 20 gallons per minute (gpm). Obviously, with 6 feet of static head already, we are not going to achieve that. Let’s see what halving the flow rate, to 10 gpm, will give us.

The loss in 1 1/8-inch pipe at 10 gpm is approximately 4.5 feet per 100 feet of pipe. We have 35 feet of pipe plus five elbows of seven feet equivalent pipe each, a total of 70 feet of equivalent pipe. Using those numbers:Static head6.00 feetFriction head70 feet @ 4.5/1003.15 feetTotal

9.15 feet

This was a useful guess, because our pump data shows a flow rate of nine gpm at a discharge head of 9.15 feet. The actual figure is, therefore, somewhere in between.

In this installation, then, we cannot expect the pump to operate at anywhere near its maximum rating. We will do well to get 50%, or 600 gph.

What can we do to improve this?

There is not much we can do about the 6 feet of static head, but we can reduce the friction head by shortening the discharge and perhaps going up in pipe size. Can we smooth out the run, losing a couple of elbows? Does the discharge have to be in the transom? It might go straight up and out the side, a distance of 15 feet. The equation will change a little when the boat heels, but not to any serious amount, especially when compared to the benefit of eliminating the equivalent of 29 feet of pipe.

Let’s use 1 1/4-inch pipe, and guess at 14 gpm, giving a resistance of 3.5 feet/100 feet.Static head6.0 feetFriction head:36 feet of 1 /14-inch pipe (15 feet + 3 elbows) @ 14 gpm1.3 feetTotal7.3ft

Another good guess. According to the chart, our pump delivers about 13 gpm at 7.3 feet of head. So, by smoothing out the run and increasing pipe size, our bilge pump now operates at nearly 50% higher capacity.

This is a dramatic illustration of how carefully planned plumbing can bring out the best possible performance from a centrifugal pump. Or, conversely, how ill-designed plumbing can hobble a perfectly good piece of equipment.

In the next example, a pump for circulating raw water to cool a refrigeration compressor is mounted adjacent to the compressor and 0.5 feet below the waterline. There is an inlet strainer, 15 feet of inlet piping with three elbows and 20 feet of outlet piping with eight elbows (including the heat exchanger) exiting via a loop three feet above the waterline. A flow rate of six gpm is specified for the compressor. The pump installed is a flexible impeller type which provides 6.3 gpm at 10 feet of head and because it has 1/2-inch pipe thread terminals, 1/2-inch hose has been used.

We need to know the total dynamic column at 6.3 gpm using 1/2-inch pipe.

In this case, the pump is below the supply, the supply being the sea, giving us a negative static lift.static lift-0.5 feetlift resistance, 15 feet pipe = 15 feet pipe3 elbows = 3 x 6 ft = 18 feet pipe1 strainer = 5 feet 5 feet pipetotal equivalent pipe = 38 feet pipeFrom chart, friction loss at 6.3 gpm in 1/2-inch pipe is 54 feet per 100 feet pipeEquivalent lift = 38 x 54/100 = 20.5 ftDynamic Lift 20.0 feetstatic head 0.5 feet + 3.0 ft = 3.5 feethead resistance, 20 feet pipe = 20 feet pipe8 elbows = 8 x 6 = 48 feet pipeTotal equivalent pipe = 68 feet pipeEquivalent head = 68 x 54/100 = 36.7 feetDynamic Head 40.2 feetTotal dynamic column = 60.2 feet

This pump doesn’t have a prayer. It is described as self-priming to four feet of lift. In this case, it will always be primed because it is below the waterline, but with 20 feet of dynamic lift, it will be struggling to draw anything at all through the supply piping. With another 40 feet of dynamic head it is doomed.

Even if this system were replumbed with 3/4-inch piping (at 15 feet loss per 100 feet) it still would not work satisfactorily. The dynamic lift would be reduced to 5.2 feet, which is close to workable, and dynamic head to 13.7 feet. But this still gives us a total dynamic column of 18.9 feet, well above the rated head for the pump.

What to do?

Space and structural constraints being what they are, certain components, such as the intake sea-cock and the compressor cannot be moved. The pump, though relatively small, should be located closer to the intake. This is always a good move as pumps are most sensitive on the inlet side. The plumbing should certainly be increased in size throughout, shortened if possible, and elbows eliminated.

By careful placement and good plumbing, the dynamic lift can be reduced to zero, especially if the pump gets lowered, giving greater negative static lift to balance friction losses. We are then concerned with the discharge plumbing and dynamic head.

The target to shoot for is 10 feet. At a flow rate of 6.3 gpm in 3/4-inch pipe, that gives us an allowance of 100/15 (the inverted resistance, giving 100 feet per 15 feet of head) multiplied by our target of 10. We have, therefore an allowance of 67 ft of equivalent pipe from pump outlet to discharge. Elbows consume this allowance rapidly and are to be avoided.

If the pump can be mounted so that it is always below the waterline, even when the vessel rolls or heels, a centrifugal pump could be used, which has the advantage of quiet operation. Having done the best we are able with the plumbing run, we can select a pump which will deliver the required volume at the discharge head.

Note: A well known manufacturer of refrigeration units used to include the raw water cooling pump tidily mounted on the same frame as the compressor. They no longer follow this practice as they had no control over the plumbing arrangements, and were plagued with pump failures as a consequence. The foregoing example demonstrates why.Appropriate uses

Improved knowledge of pumps and the quirks of associated plumbing will lead to a better understanding of what types of pumps are suited to what applications.

Reciprocating pumps, frequently of the multi-chamber diaphragm type, are used in pressurized fresh water systems where their ability to run dry allows them to survive an empty tank. Coupled to a pressure switch and an accumulator (to smooth out the flow), they are suited to the intermittent use encountered in this application.

Ease of inspection and maintenance favors the flexible impeller type of rotary pump for hard-working environments such as the supply of cooling water to an engine or power plant.

Silent operation and adaptability to varying flow demands makes the centrifugal pump ideal for circulating water through air-conditioning systems.

All of these pumps, though, have their limitations. If they are to give long and satisfactory service, they need to be properly matched to their service and protected from damage through the depredations of wear, foreign matter, misuse, and bad plumbing.

In order to minimize the potential for problems down the line, it is crucial at all times to follow the recommendations of the manufacturer of any pump with respect to appropriate use and installation. If the information supplied provides insufficient guidance, demand more. Be fully informed about the product, or don’t use it. As we have seen, problems with pumps do not always have obvious answers. It makes no sense to start out in doubt.

Jeremy McGeary is a naval architect who lives in Newport, R.I.

By Ocean Navigator