Pumps & Systems, June 2007
Last month, we discussed applications of low head axial flow pumps ("Field Testing: Theory vs. Reality," Pumping Prescriptions, May 2007). Knowing your system is very important for these types of units. For example, consider a recent application where neglecting the system caused problems.
Our customer was in the process of designing a system where liquid in the storage tank had to be recirculated by the pump in order to keep it from polymerization and solids settling. The requirements given to our applications group were 800-gpm flow and 10-ft head - a perfect application for a 6-in axial flow pump unit. Two identical units were requested. We entered the parameters into our computer to predict 73 percent efficiency and 2.8-hp requirements, as shown in Table 1. A 3-hp motor was selected.
Table 1. Centrifugal pump efficiency estimator, Courtesy of Pumping Machinery:
The required delivery was tight (two weeks), with little room for error. The units were shipped on time and quickly installed by the customer. The next day, we received a call from the customer: "Guys, we have a problem. One of the units barely produces the required 800-gpm flow, while the second one overfloods the tank and also trips the motor." Needless to say, we got onsite right away, and discovered the following:
Figure 2. (A) The unit does not get enough flow (left). (B) Too much flow, and the motor trips (right).
Our pump was supposed to have 6-in piping between the pump and the tank, with negligible friction losses (see Figure 3), based on the Cameron Hydraulic Data Book.1
Figure 3. For Schedule 40 6-in pipe, friction losses at 800-gpm are only 4.04-ft per 100-ft pipe, i.e. 1.8-ft for the 45-ft of 6-in pipe run we have here.
The system should have been mainly a "static head" system (10-ft), allowing 800-gpm, with 2.8-hp motor draw. Unfortunately, as the maintenance manager explained, even though they actually installed unit (B) first with the proper 6-in pipe, they eventually ran out of pipe and finished up unit (A) with a 4-in pipe. Bingo! That explains the reason for low flow on unit (A) (see Figure 4).
Figure 4. A smaller pipe exerts significantly more friction losses, which increase pump head and choke off flow.
Note the differences in piping between (A) and (B). Since the pipe on unit (A) is not submerged under the water level, the pump needs to always push against 25 - 15 = 10-ft static head (plus friction). Unit (B), however, is different: the discharge pipe is submerged under the water level, thus the resultant pump head is actually zero (or very small due to some small friction head in a 6-in pipe):
Figure 5. Unit (B) has the discharge pipe submerged under water, creating an entirely different resultant pump head.
System (B), with much lower static head, allows the pump to run way out in flow, drawing significantly more horsepower and overloading the 3-hp motor.
The corrective action was then obvious, to make sure that the pipe is 6-in on both units and static head is the same on both as well: 10-ft. With the systems corrected, both produced the required 800-gpm, with motors running smooth and cool, as they should.
As always, a Pump School Quiz: While this corrective action worked, the ultimate system still should not be considered optimal. Can you tell why? The first correct answer wins you a ticket to our next Pump School training session.
1. Cameron Hydraulic Data Book, Flowserve Corporation, edited by C.C. Heald, Nineteenth Edition, First Printing,