When selecting an electric motor it is important to select a motor that can develop enough horsepower to drive the pump. Often motors are selected to be non-overloading at the end of curve. This means that even at the maximum power requirement for the pump, the motor is large enough to drive the pump without overloading. In submersible pumps there are several important factors that come into play for proper selection. We will start from the top of the motor and work our way down.
Obviously if the motor is submerged in the liquid the power cable must enter the motor housing at a junction box that is below the liquid level. This is a prime location for a leak. Different manufacturers have different philosophies regarding how this connection should be made. Many believe that this should be a rigid permanent connection with built in strain relief. This often has a packing gland around the entrance of the junction box and occasionally a secondary seal to prevent leakage. Other companies see an advantage to having a quick disconnect on the cable allowing the pump to be replaced without the need to re-cable the unit to the Motor Control Center (MCC). If pump change out is frequent because of the need to de-rag the pump or other operational problems, it may be advantageous to have a pump with a quick disconnect. This may be especially true if it allows the unit to be changes without the need for calling out an electrician.
Generally in submersible pump designs there are two bearings to be concerned with. The upper bearing that generally designed to support the rotor (pump impeller, shaft, and motor rotor) in the radial direction. This bearing is allowed to move axially, via a slip fit to the housing. This allows for thermal growth in the rotor as it heats during operation. The lower bearing which is generally responsible for supporting the rotor in both radial and axial loading is fixed in place allowing it to transfer the axial loads from the pump into the motor frame.
For pure radial loading in pumps, generally a single row deep groove ball bearing (A) works well. They are inexpensive and have more than enough radial load capability for common pump applications.Occasionally double row deep groove bearings (B) and roller bearings (E) are used when very high radial loading is expected but these are often more expensive, and are generally found on large horizontal pumps.
For the combination of radial and axial loading the single row, deep groove bearing can be used but the axial load capability is somewhat limited. The double row, deep groove bearing supports much higher axial loading (often 1.7 to 2 times the capability). When very high axial loads are expected angular contact (C) and tapered roller (D) bearings can be used that have 3 to 5 times the load carrying ability but can only do so in one direction.
A caution regarding bearing sizes: often in specifications I see mean time between failures (MTBF) or mean time between repairs (MTBR) specified at very high numbers (60K, 80K, 100K hrs). This leads the manufacturers to install larger bearings to meet the very high design life requirements. Bearings running in very lightly loaded conditions often cause the bearings rolling elements to skid instead of roll. The skidding action causes the lubricant film between the rotating element and the raceway to dissipate because more oil is not being drawn in by the rolling action of the bearing eventually leading to metal to metal contact. This leads to spalling of the bearing and eventually to its failure. In my career I have seen many more bearings that have failed due to under-loading than over-loading. The most common cause of bearing failure is actually improper lubrication either from contamination of the lubricant, or poor preventative maintenance practices.
Oil Filled vs. Air Filled Motors:
Oil filled motors offer several benefits, the most pronounced being that because of the much higher thermal transfer capacity of oil vs. air (approximately 7X) oil filled motors tend to run cooler. The oil also provides continuous lubricant for the bearings and the windings. It is claimed by some manufacturers that the vibration or start up torque pulses of the windings causes the insulation to wear eventually leading to shorts within the motors. Oil filled motors are supposed to lubricate the windings and prevent degradation from chaffing during start-up. I don’t see this as a big problem unless you are putting an excessive amount of starts and stops on your equipment. There are also studies that claim that oil filled motors prevent moisture from getting into the insulation on the windings. The insulation is hydroscopic and tends to breakdown more quickly in moist environments.
Air filled proponents tend to focus on the fact that there is a higher amount of drag loss in an oil filled motor compared to an air filled design. Typical estimates range from 1% to 2% more loss. They talk about the need to periodically replace the oil, but from my research this is only recommended if the mechanical seals fail and product gets into the bearings. I believe in an air filled motor the grease in a grease filled bearing would also need to be replaced after a seal failure, so I will call that one a tie. Air filled supporters also expound about the possible catastrophic affects of an oil leak from the motor we must keep in mind that we are not talking about the Exxon Valdez. This is a small amount of oil and generally these days, manufacturers use non toxic blends of oil. It all really comes down to efficiency and heat dissipation. If the liquids are always cool and provide plenty of heat dissipation, an air filled motor will probably work just fine. If heat dissipation might be an issue, then I would look very closely at an oil filled design.
If you plan to use your submersible pump in a dry pit application heat dissipation is a major concern. Many manufacturers will require that the motor stator be jacketed to help remove the motor heat. The jacket circulates liquid over the outside of the motor helping to dissipate the heat. These generally come in two forms, product cooled and self contained. Product cooling passes some or all of the product being pumped through the jacket passing the motor heat into the pumped liquid much like it would be in a submerged application. The problem with this design is that large particulate may cause plugging of the jacket ports and lead to a motor failure. Various port designs and configurations have been developed to prevent this, but a frank conversation with your pump vendor should happen before you purchase this option. The self contained option uses a separate cooling loop to pass clean liquid over the motor and generally does not suffer from the plugging problem.
The job of the mechanical seal is to prevent the liquid you are pumping from leaking up into the bearing and motor housing. This is accomplished by rotating one extremely flat seal face in very close proximity to a stationary face of approximately equal flatness. The faces are lapped flat to within 2-4 helium light bands of being perfectly flat. In the space shown as the fluid wedge in the diagram below, a small amount of liquid is wicked through the faces and is vaporized by the heat generated by the faces rotation. There must be liquid in the seal faces to cool and lubricate them. When seal faces run dry they fail quickly, sometimes within fractions of a second. In a typical single seal pump it is the pumped liquid that passes through the faces and is vaporized as it enters that atmosphere. In a submersible pump some small units use a single seal to seal the motor from the pumped liquid.
Figure 12: Typical single seal in a process pump
In most submersible pumps the typical seal arrangement is a dual seal in either a double or tandem arrangement.
The dual seal offers the protection of two seals to prevent the pumped product from getting into the bearings and motor. The tandem arrangement has both seals facing the same direction. The bottom seal is in the pumped liquid (A), as the pressure increases in the pumped liquid it actually forces the seal faces closer together reducing the amount of liquid that passes through the faces. The oil in the seal chamber (B) is intended to be the lubricant for this lower seal. The upper seal is feed by the oil in the motor or bearing housing.
Figure 13: Typical tandem dual seal in a submersible pump
With a double seal the two seals are positioned back to back. This has the advantage that both seals are operating in the clean oil environment. The disadvantage is that a pressure spike in the liquid on the pump side of the seal can cause the seal faces to force open and product can be introduced into the seal cavity. This reduces or destroys the lubricity of the oil in the seal chamber, eventually leading to a failure of the bearings.
Seal Face Materials:
Seal faces can be made from a variety of seal materials. The most common materials are carbon (F), ceramic (C), tungsten carbide (WC) and silicon carbide(SiC). Carbon is a good seal material because it somewhat self lubricating and is fairly inexpensive. Its nemesis is abrasives. It is such a soft material that it is easily scratched. That scratch provides a leak path and the seal fails. Ceramic is often paired with carbon. It is harder than carbon but generally not harder than common abrasives. It too is easily scratched leading to failures in abrasive environments. It also does not have the mechanical strength of Tungsten and Silicon Carbide (it flexes under pressure). It is susceptible to thermal shock (quick temperature change), causing it to shatter. Its primary attribute is that it is inexpensive, and therefore very popular. Tungsten Carbide is an extremely hard material that has very good mechanical properties coupled with excellent corrosion resistance. The material does extremely well in abrasive services. Silicon Carbide, like Tungsten Carbide is extremely hard even slightly harder than tungsten carbide. It has excellent corrosion resistance and very good mechanical properties. It is often paired with tungsten carbide as a combination of faces in abrasive services.
In most wastewater applications it is desirable to have primary seal made from harder materials so I would look for SiC vs SiC or SiC vs. WC. The upper seal should only be pumping oil so a carbon ceramic seal would be a good choice. Some companies offer both seals with hard faces but unless there is a particular reason, I would use F vs. C to reduce the amount of heat generated by the seal and reduce the cost of the pump.
Some manufacturers use proprietary seal designs and proprietary seal materials. The debate as to whether the proprietary seals are better than the seal commercially available by the seal manufacturers seems to be based more on opinion than fact. I always fear a single source for a component because it often leads to higher prices on repair parts and long possible lead times. If you don’t mind funding your own insurance policy by tying up your working capital up in you spare parts inventory then proprietary seals may be a good option for you.
So what if the mechanical seals fail? How would you know? On typical process equipment the puddle under the pump is your first warning. With submersible equipment there is no such visual indicator. Most manufacturers of submersible equipment have at least one moisture detection device in the pump. I will not attempt to explain all of the variations in technology that this moisture detection sensor may employ. Most are time tested technology that will give you an indication that you have moisture in a portion of the pump that should not. I do believe the location of these sensors is important. Some manufacturers mount the sensor in the cavity between the primary seal and the secondary seal. If you primary seal leaks the secondary seal prevents the liquid from getting to the bearings and motor where it can cause damage. Most Manufacturers using this method additionally state that the repair of the pump can then be scheduled for a convenient time, the pump does not need to be shut down immediately (just very soon). I believe that this is the proper position for such a device. Other manufactures locate the sensor in the motor cavity which means that you already have liquid in the bearings and the motor when the sensor alerts you that you have a problem. This would then require an immediate shutdown of the unit to prevent further damage. This warning seems a little late in my opinion, (Like quitting smoking after you have N-stage lung cancer).