Don’t ‘Over-FEED’ Your Pumps (Part 2)

Right-Size Your Pump System by Leveraging the Collective Experience of Engineering and Operations Personnel for More Accurate Design Specifications 

By Ray Hardee

The Story: Your Accumulated Knowledge Can Prevent Over-Sizing 

Our goal is to minimize the number of assumptions that we make during the FEED process, and to reduce the magnitude of the design margins and avoid over-sizing equipment.  

In this past blog post, we discussed the FEED (Front End Engineering Design) process in pump specification and saw how it can lead to over-sizing pumps—a practice that can result in equipment not operating efficiently within the system.

We noted that engineers often include a safety factor to account for system unknowns during the FEED process as “insurance” for the system design. I don’t like the term “safety factor,” because all engineering designs should be safe.  I prefer the term design margins to describe how we account for uncertainties in the preliminary design process.

This month we’ll follow the FEED process from my last article, but we’ll use will use the concept of design margins to document these assumptions and share them with everyone involved in the process.  Our objective is to use the team’s collective experience to gain an understanding of how the system will operate, based on the design and operating history of similar piping systems. Our goal is to minimize the number of assumptions that we make during the FEED process, and to reduce the magnitude of the design margins and avoid over-sizing equipment.  

We’ll also conduct a system review to help identify concerns that could occur after the system is placed in operation.  Let’s get started, and see how to FEED your pumps right.


The Key: Knowing How the System Will Really Operate


All piping systems consist of pump elements, process elements, and control elements working together to meet each system’s operating requirements.  The pump is sized to pass the design flow rate through the process and control elements to make the product or provide the service. 

In our example, the operations team at a process company is planning for an upcoming facility expansion.  After completing a market study, the team members determine that the system needs a capacity of 600 gpm for the foreseeable future.  Based on their experiences from similar projects, they realize that the system may run at a lower capacity for the first year or two.  But the company is very confident in its product, and wants to design the system with the flexibility to boost its capacity in the future—if  that could be done at a reasonable cost. 

They engage the services of an engineering firm and define the system’s capacity requirements of 600 gpm and document their experiences in initially operating the system at a much lower flow rate for up to two years.  They also establish the goal of expanding production capacity to 1,200 gpm to meet potential market demand.   

The design firm comes up with two options for consideration. The first consists of sizing the system to meet current conditions, along with the proposed future design capacity of 1,200 gpm. Option 2 consists of sizing the system for the current requirement of 600 gpm, with the ability to add a second 600 gpm parallel stream in the future. Initially, the second option is believed to be the less-costly alternative.

After further evaluation, however, it’s discovered that the cost for a heat exchanger sized for 1,200 gpm is less that the cost of two heat exchangers sized for 600 gpm. Furthermore, the cost of a single pump for 1,200 gpm costs less than two pumps sized for 600 gpm. This information makes Option 1 begin to look more favorable.  

The company’s operational procedures require installation of a standby pump to prevent any process system from shutting down after the failure of a single pump. This mandates the purchase of a standby device for both options. Initially, when operating at the expected 600 gpm flow rate, the cost for the standby pump favors Option 2.  Yet, upon considering the proposed future capacity, the cost of two larger pumps for Option 1 is similar to the cost of three smaller pumps required by Option 2.  

Based on these and other considerations, the company decides to proceed with Option 1.  Additionally, since the expected flow rate through the system varies greatly, it decided to incorporate a variable frequency drive (VFD) to enable the pump to better meet system requirements. This results in the same flow sheet presented in our last Pumps & Systems article, but now everyone has a clearer understanding of how the system will operate over its proposed lifespan.  

Image 1 – Flow diagram showing the locations and elevations of the equipment along with the details needed for pump sizing.  



Using Design Margins vs. ‘Safety Factors’

During the FEED process, we want to use the collective experiences of the engineering team to gain a better understanding of how the system will operate based on the design and operating histories of similar systems.  This should minimize the number of assumptions made during the process.  To demonstrate, Table 1 compares the pump sizing calculations using the safety factor method (presented in last month’s article) and the design margin method.  

The engineering firm creates a set of informed assumptions for the system’s process and control elements based on the firm’s sizing criteria.  For example:

  • Static Head:  Under normal operations, Tank 1 has a level of 5 feet, and Tank 2 has a normal level of 5 feet and a tank pressure of 15 psi.  This results in a calculated static head of 84.6 feet (the same as the past article) utilizing a safety factor (See Table 1, below).

During plant operations, variations in tank levels and pressures must be accounted for in the design.  Tank 1 is open to the atmosphere and can operate at a low tank level of 1 foot, resulting in a minimum tank energy value of 101 feet.  Tank 2 is a closed tank and has a maximum operating pressure of 20 psi and a maximum operating level of 9 feet, resulting in a maximum tank energy of 206.2 feet of fluid.  The resulting maximum possible static head is 105.2 feet of fluid. The system may operate like this for less than 5% of the time, but the pump must be sized to meet this extreme requirement.   

  • Pipeline Head Loss:  We’ll start our design by using an as-built pipe list from a previously completed project of similar design.  We have realistic estimates for pipe length, along with the count of isolation and check valves, tank penetrations, and elbows needed for the head loss calculations.  Because the FEED stage occurs before the actual pipe routing, we need to assign design margins for the unknown values associated with pipe length and elbow count.  

The engineering firm’s pipe specification document provides design margin guidance when calculating the pump’s design point head.  The as-build estimate for pipe length is increased by 20%, and the number of elbows is doubled when performing pump sizing head loss calculations.  This results in a pipeline head loss of 51.4 feet.   

  • Process Element Heat Exchanger (HX): The firm’s heat exchanger specification states that the maximum allowable differential pressure shall be no greater than 15 psid at the design flow rate.  A design margin of 5 psid pressure is added to account for heat exchange tube fouling.  This results in a heat exchange head loss for pump sizing of 46 feet of fluid.  Under normal operation, the heat exchanger head loss will be less that this value, but the value is used for pump selection calculations. 
  • Control Element Valve (CV): The control valve specification states all valves should be sized with a differential pressure of 15 psi with the valve in the 80 percent open position at the design flow rate.   NOTE:  The operating company decided to leave the control valve as a backup, in the event the VFD would not be able to maintain the flow rate over the range of operations.

Table 1 compares the “safety factor” and “design margin” methods when arriving at a pump selection design point.  In both methods, the pump head equals the sum of the calculated head loss for the process and control elements at the design flow rate.  

Remember: in the safety factor method, each reviewer adds their own safety factor based on their work experiences in their area of expertise.  It’s difficult to determine the total safety factor, but in last month’s example, we proposed a 30% safety factor.     


Pump Sizing Based on Safety Factor and Design Margin Methods 


Values using safety factor

Values with design margin

Static Head 

84.6 feet

105.2 feet

PE Pipe 1-4 Head Loss

56.5 feet

51.4 feet

Process Element HX

34.6 feet

46 feet (including tube fouling)

Control Element CV

34.6 feet

35 feet

Calculated Pump Head

210.3 feet

225.4 feet

Safety Factor

30% of calculated pump head

No overall safety factor

Head value in pump spec



Table 1 – The method of determining the pump head sizing requirements comparing the used of “safety factor” and “design guides”.

Using the design margin approach, the number of design variables have been reduced by:

  • Arriving at a better understanding of the system flow rates during the system’s lifetime
  • Calculating the maximum static head based on expected plant operations
  • Using as-build data from previous system designs to arrive at starting estimates for pipe length and routing
  • Identifying the uncertainties in pipe length and number of elbows, and factoring in a documented design margin
  • Specifying a realistic value for heat exchanger fouling based on plant operating experience
  • Factoring in the pressure drop across a control valve at the design flow rate.

The design margin values are specified in each step of the pump sizing calculations. When questions arise, knowing the design margins, and how they’re determined, gives everyone a clear view of the process.

Based on the results of the FEED calculations, the engineering firm provides a pump specification with a design point of 1,200 gpm and 226 feet of head.  


The Pump Selection Process 

The pump specification, complete with the design point and the requirement for a variable flow drive (VFD), is sent to each supplier on the pump bidders’ list.  

Because the pump design point is the maximum value expected for the system, many of the suppliers now recommended a pump with the best efficiency point flow closer to the 1,200 gpm. The pump suppliers still increase the impeller diameter to take better advantage of the selected pump motor size; some habits are hard to break.  With the use of a VFD as the control element, the flow rate through the system can be adjusted by changing the pump’s impeller speed to meet actual system requirements.    

The successful pump supplier recommended a 6×5/16 end suction pump with a 15 1/8-inch diameter impeller operating at 1750 rpm consuming 88 hp.  The pump’s BEP of 75% occurs at 1,000 gpm, and the pump is supplied with a 125 hp motor with a VFD.

The engineering firm reviews the pump recommendations, accepts the bid, and creates a purchase order.  The supplier accepts the PO, sends the document package for use by the entire design team, and then sends the pump to the jobsite.    


Placing the System into Operation 

Once the system is complete and turned over to operations, the current market only requires 400 gpm.   

A system energy balance is developed using the manufacturer’s supplied pump curve and heat exchanger pressure drop data.  Energy balances are performed at the current operation of 400 gpm, the expected operation of 600 gpm, and the future design operation of 1,200 gpm.  

The system was controlling successfully using the VFD.  Remember, the control valve was only included as a backup if the VFD failed to work. The head loss across the fully open control valve is factored as a control loss and could be thought as backup insurance for the VFD.  

The pipeline head loss was calculated based on the installed system piping, and the actual static head calculated by the normal operating difference between the tanks. The heat exchanger values were based on the manufacturer’s supplied data without fouling.  



Current Operation

400 gpm

Expected Operation

600 gpm

Future Operation

1,200 gpm

Pump Element head

98.2 feet

114.8 feet

191.6 feet


Process Elements


Static Head 

84.6 feet

84.6 feet

84.6 feet

Pipeline Head Loss 

5.7 feet

12.5 feet

49.3 feet

Process Element HX

3.9 feet

8.7 feet

34.6 feet


Control Element 

4.0 feet

8.9 feet

23.1 feet

Table 2 – The system energy balance showing how the energy is used.

Table 2 shows how the energy is used in the system.  Notice that the energy consumed in the process elements and control elements equals the energy supplied by the pump.  By using a variable speed drive on the pump element, the pump speed only provides the actual energy required by the process and control elements.  



Current Operation

400 gpm 

Expected Operation

600 gpm 

Future Operation

1,200 gpm 

Pump Element


Pump Speed rpm 

1094 rpm

1208 rpm

1691 rpm

Pump BEP Flow 




Percent of BEP Flow




Pump efficiency (%)




Pump power (HP)




Table 3 – How the pump with the variable speed drives works through this large range of flows.  

In looking at Table 3, please note that a change in the pump speed directly affects the best efficiency point, along with the percent of BEP flow. Also notice that the pump power has a dramatic decrease at the lower flow rates.  This is attributed to the fact that, at lower flow rates, the head loss in the process element decreases, which allows the pump to operate at a lower speed.   

The “Rotodynamic Pump Guidelines for Operating Regions (ANSI/HI 9.6.3)” standard recommends operating the pump with the flow rate between 80% to 110% of the BEP flow.  Notice the pump with the VFD has the percent of BEP flow closer to the actual flow rate through the pump, and the pump operating around 600 gpm is in the middle of the sweet spot.  


Conclusion: Over-Sizing Pumps can be Avoided 

In the series of articles about the FEED process for pump selection, the design flow rate and total developed head requirements are key values for a pump specification.  Since pumps are such long-lead items, these calculations must be completed well before many of the system details are available. This results in making assumptions on the final design.  In the January article, a “safety factor” was added during the calculations and their review.

This month, we covered the design margin method, with a focus on understanding how the system will operate and documenting the process. These systems live a long time, and their initial design can be modified multiple times.  We saw how team members used their operating experiences in developing a long-term project plan. 

By providing their engineering firm and equipment suppliers with this long-term plan, each group can incorporate flexibility in its initial design, while incorporating methods to minimize the expenses in the future.  

We also discussed the value of using previous projects of similar design as the starting point for the FEED process.  Design margins were also spelled out and used to realistically identify items that could vary during the intermediate and final design process.  By choosing them based on the combined experiences of all the groups involved, we reduced the probability of over-sizing equipment.

Leave a Comment

PIPE-FLO Newsletter

Stay in the KNO about PIPE-FLO

A software pipe flow program which helps on how to measure pressure in a pipe

4529 Intelco Loop SE
Lacey, WA 98503


Copyright © 2020 Design by PIPE-FLO