Chementator Briefs
Polyesters Customizing the macroscale properties of polyesters by using sidechain…

Comment PDF

An Overview of Vacuum System Design

By Patrick Govoni |

The design of vacuum systems deserves careful attention — there are multiple facets that affect efficiency, operability and cost

1Understanding vacuum and designing vacuum systems can be difficult, and there are many unique aspects that an engineer must take into account. Although some aspects of compressed-gas system design can be translated to vacuum-system design, for the most part they are two entirely separate subjects. This article discusses common misconceptions, such as units of measurement and calculation of flow. It also covers methods of producing vacuum and design aspects, such as minimizing air leakage, pressure drop, condensation and a specific approach to sizing vacuum piping. The design of vacuum systems is certainly a topic that deserves careful attention given that there are multiple facets that affect efficiency, operability and cost. After a few key changes in perspective, the reader will find the subject is not as intimidating as it first appears.


Introductory concepts

Defining vacuum. The common definition describes vacuum as a space void of matter. However, at the current level of technology, scientists have not demonstrated that it is possible to remove all gas from a given space. The practical definition would be that vacuum is a space where the pressure is lower than atmospheric pressure. In casual conversation, the definition of vacuum is often incorrectly defined as the pressure differential or the force created between an area of lower pressure and an area of higher pressure. Whichever way it is defined, the functions and applications of vacuum are myriad and span across industries. The core applications include removing active atmospheric components that could cause physical or chemical reactions for dairy packaging applications, achieving a pressure difference for lifting, reducing heat or electrical energy transfer for insulation, or removing dissolved gas or volatile liquid from materials for freeze drying.

Units. Understanding units is the first challenge encountered when learning about vacuum. For most common pressure units, a particular pressure lower than atmospheric can be described in three distinct ways: an absolute pressure (4.7 psia), a gage pressure (–10 psig), and a vacuum pressure (10 psi vacuum). This ambiguity is prone to cause headaches, especially if more than one of these units are used in the same conversation or document. The torr unit (1 torr = 1 mm Hg) is therefore often preferable, because it is defined on an absolute scale and eliminates any ambiguity. The most common pressure units and their relative values are shown in Figure 1.

Figure 1. This graphic shows the different units one can use when discussing vacuum [1]

Figure 1. This graphic shows the different units one can use when discussing vacuum [1]

Flowrates. Flowrate is another main area that may confuse engineers new to the topic, namely the relationship between standard volumetric flow in standard cubic feet per minute ( SCFM), pressure, and actual volumetric flow at the system pressure in actual cubic feet per minute ( ACFM), as given by Equation (1):

Print  (1)



 (Note: all parameters are defined in the Nomenclature box above).

For multiple-component gas mixtures, a slightly different equation must be used since condensable vapors, such as water, can significantly alter flow. This is shown in Equation (2) for a two-component mixture:

 Print  (2)



 Higher-pressure gas entering a lower-pressure piping system expands as a result of the pressure difference, thus ACFM will be larger than SCFM (Figure 2). But at the lower end of vacuum pressures, it becomes apparent this relationship is exponential. One standard volume of gas expands 7.6 times upon entry to a 100 torr system, 15.2 times to a 50 torr system, and 76 times to a 10 torr system. By contrast, in compressed gas systems, the pressure term changes less than 0.06 for every change of 50 torr. Consequently, this is why vacuum piping is often much larger than compressed-gas piping, even if flow demand is lower for the former.

vacuum system design

Figure 2. Shown here is the gas expansion that takes place upon entry into piping under vacuum

To account for this volumetric sensitivity to pressure, it is helpful to view flow using throughput, qPV, as given in Equation (3):

Print  (3)



 Throughput, which is equal in value at all points of a closed system, is generally used to illustrate overall mass flow through a system. The addition of the pressure term adds context to the volumetric flow and often provides a better representation of system demand than a simple volumetric flowrate. However, the temperature must be constant throughout the system for throughput and mass flow to be related, unless one wishes to use significantly more complicated adiabatic flow equations. To understand flow at certain points in a system at a particular pressure, conductance, C, is used, as shown in Equation (4):


 Print (4)



As with electrical conductance, vacuum conductance is a reciprocal of flow resistance and indicates the ability to allow passage. Resistance in vacuum flow is caused by friction between gas molecules and the wall surface and friction between the gas molecules themselves, resulting in pressure differences and volumetric flowrate losses. Conductivities can also be connected by following the same rule as its electrical counterpart, adding individual conductivities when in parallel and adding the reciprocals of the conductivities, or resistances, when in series.

Pumpdown. Implementation of vacuum principles is very different based on application, such as whether the system flows or is closed. Where filtering or drying demands a constant flow of vacuum, evacuating and maintaining vacuum in an enclosed space demands a lower volumetric flowrate as the ultimate vacuum pressure, a parameter determined by the efficiency of the vacuum pump, is approached.

The ultimate vacuum pressure of a vacuum pump is specific to the design of that particular pump and depends on characteristics such as the vapor pressure of the oil or other sealing liquid and the degree to which the system leaks. The logarithmic rate at which a vacuum pump approaches its ultimate vacuum pressure, however, can be explained by a simple concept that applies to all types of mechanical vacuum pumps. The density of a vapor decreases as pressure decreases, and mechanical vacuum pumps are constant-volume devices. Thus, the number of molecules that are displaced in each successive volume is gradually less as a vacuum pump reaches the lower-pressure regime.

A simple illustration of this phenomenon would be a vacuum pump curve that shows a decrease in inlet capacity as inlet pressure decreases until the ultimate vacuum pressure is reached, such as shown in Figure 3. An estimate of the time required to pump down from a specific pressure, p 0, to another pressure, p 1, is given by Equation (5):

Figure 3. This graph shows the relationship between pumping speed and pressure of a typical rotary-vane vacuum pump [2]

Figure 3. This graph shows the relationship between pumping speed and pressure of a typical rotary-vane vacuum pump [2]


Print   (5)



However, this time estimate is often shown to be consistently lower than the actual evacuation time due to leakages and pump inefficiencies.


Vacuum pumps

There are several types of vacuum pumps and each type has its own benefits and drawbacks that must be considered when making the best choice for a specific application. A vacuum pump is essentially an air compressor operating in reverse, where it compresses the air or gas in the vacuum system and discharges it into a vent.

Diaphragm pumps. The simplest vacuum pump is a diaphragm pump, which consists of an inlet valve, outlet valve, and diaphragm that sucks in gas on expansion and pushes gas out on contraction (Figure 4). Mostly used for laboratory benchtop or other small applications, diaphragm pumps are oil-less and water-less, with the pump mechanism sealed off from process fluids. These are best for low-flow applications (pressure greater than 100 torr) and for those with a fluid with contamination or other chemical sensitivities.

Figure 4. This diagram illustrates the operating principle of a diaphragm vacuum pump

Figure 4. This diagram illustrates the operating principle of a diaphragm vacuum pump

Rotary-vane pumps. Rotary-vane pumps are used in many different situations and can handle pressures below 10–3 torr. A rotary vane pump (Figure 5) consists of a single eccentric rotor with vanes inside a larger cavity. During suction, a vane brings in the gas, rotating until the other vane closes off a volume of gas from the vacuum system. There is further compression as the vanes rotate, eventually expelling the gas through the outlet valve.

Figure 5. This drawing shows the operating principle of a rotary-vane vacuum pump [3]

Figure 5. This drawing shows the operating principle of a rotary-vane vacuum pump [3]

This type of pump requires lubrication oil that may require an oil separator for exhaust. If the pump is pulling potential condensables, a gas ballast can be included that reduces the pressure needed to open the outlet valve, thus reducing the likelihood of vapors condensing in the pump cavity. A high volume of condensables, especially solvents, can negatively impact these pumps, as they have a tendency to mix with the seal oil and can corrode the pump internals if not handled properly.

Liquid-ring pumps. Liquid-ring pumps are common vacuum pumps; they attain pressures down to 25 torr, and are preferred from a reliability perspective because of their low-friction design. Similar to rotary-vane pumps, liquid-ring pumps consist of a single eccentric rotor that pulls gas with an initially expanding and later contracting cavity using a ring of liquid, usually water, that acts as the boundary between the gas volumes. Solvents such as hydrocarbons and other liquids can be used as well, depending on the application. A liquid separator on the exhaust and filtration for contaminants that become trapped in the liquid may be needed. These pumps are quite amenable to a large condensable load; while temperature loss due to condensing vapors can lower efficiency, corrosion is not an issue as it would be for oil-sealed or dry pumps.

Rotary-lobe pumps. Rotary-lobe pumps, also called Roots blowers or booster pumps, consist of two lobed rotors that spin in opposite directions in a casing with tight clearances (Figure 6). This type of pump is limited by its design, in that a high pressure differential between inlet and outlet causes a significant amount of heat generation in the rotors, which can cause contact or seizure. This attribute is why a rotary lobe pump does not exhaust directly to atmosphere and is often paired with a mechanical backing pump. Although an overflow valve can be added to prevent this heat build-up, its addition would further limit the maximum possible pressure differential of the pump, making the backing pump even more necessary.

Figure 6. This illustration shows the operating principle of a rotary-lobe vacuum pump [4]

Figure 6. This illustration shows the operating principle of a rotary-lobe vacuum pump [4]

However, the maximum pumping speed of a rotary-lobe pump is limited to pressures between about 50 and 0.1 torr, which is where other mechanical pumps typically start to slow down and where water vaporizes at the highest rate. This is quite convenient on both counts, as water can add an excessive workload and increase pumpdown time considerably on single-pump systems. Rotary-lobe pumps are often paired with a backing pump to accelerate startup to steady state in sensitive applications where speed is critical. Because these pumps are most efficient in the lower pressure regime and the potential heat build-up, these pumps are not started until the pressure has been reduced enough so that their motor is not overloaded.

Ejectors. As an alternative to a vacuum pump, a simple ejector (Figure 7) can be an elegant solution to vacuum generation. Ejectors, often called steam-jet ejectors, are inexpensive and simple to install, operate and maintain. Instead of using electricity for power, a motive gas must be supplied, usually steam or compressed gas, to create the suction force that generates the vacuum. The motive gas approaches the nozzle at a high pressure and low velocity. As it passes through the nozzle, the gas expands, dropping in pressure and increasing in velocity. The suction gas, at a lower pressure and low velocity, is drawn into the chamber via the high-velocity motive gas and is combined. The two fluids travel through the throat and begin to expand along the diffuser outlet, gradually decreasing in velocity and increasing in pressure to a slightly higher velocity and much lower pressure relative to that of the motive gas at the inlet.

Figure 7.  This cutaway diagram shows the operation of an ejector pump

Figure 7. This cutaway diagram shows the operation of an ejector pump

Ejectors can be placed in series, both to increase and widen the operating range of vacuum level or flow. Condensers can be included before, in between, or after ejectors to increase the efficiency of recapture of the motive fluid. An ejector is particularly desirable over pumps in fast-cycling applications due to its lack of moving parts and speed to establish vacuum. If a constant, high-vacuum flow is required, however, ejectors may not be as energy efficient as a pump.


Vacuum system components

Valves. Valves used in vacuum applications are differentiated mainly by their conductances and leakage rates. Gate valves have a very high conductance in that they have an unobstructed, straight-through orifice and a short distance between ends. Because of their high conductance, gate valves are often placed between the vacuum system and the pump. However, these valves have higher leakage rates than other choices, so pump capacity must take this into account.

Ball valves are inexpensive, have a fairly high conductance, and a modest leak rate. Despite these favorable parameters, for inexplicable reasons, ball valves are not often used in many vacuum applications.

Butterfly valves are mostly used for conductance control, where the travel of the valve’s disc varies between high conductance when fully open and low or zero conductance when fully closed. Depending on the application, an O-ring can be included on the disc to give zero conductance when the valve is closed or not included to give a fixed, low conductance.

Angle valves are often used on top of tanks or in other situations where space is an important consideration. These valves are simple to maintain and install, have low leak rates, but a lower conductance than similar, straight-through valves.

There are two main valves used for vacuum control: vacuum regulators and vacuum breakers (Figure 8). Vacuum regulators work by limiting decreases in vacuum level, essentially as pressure-reducing regulators in reverse. When the inlet reads a loss of vacuum, or an increase in pressure, beyond the setpoint, the valve opens. And the opening valve allows the lower pressure, downstream vacuum to restore the upstream vacuum to its original level. Vacuum breakers essentially work in an opposite fashion, by limiting increases in vacuum level.

Figure 8: These diagrams show the priciples behind a vacuum regulator and a vacuum breaker [6]

Figure 8: These diagrams show the priciples behind a vacuum regulator and a vacuum breaker [6]

Receivers. A receiver, or knock-out pot, is a tank in between a pump (or pumps) and a piping system that serves as both a vacuum pressure-stabilizing element and a liquid-catch tank. As the largest volume in a vacuum system, a receiver effectively increases the time required to both lower and raise pressure. By increasing the time required to change pressure, it provides a vacuum storage time in between pump operation and rising or falling demand from opening or closing inlets. In addition to steadying the vacuum level of the system, a well-designed receiver increases pump reliability by reducing pump short-cycling.

A receiver also provides a catch-tank for liquids to provide a barrier between vacuum generation and piping system for both condensables entering through inlets and for liquid pump sealant traveling to the distribution network. Often, such a tank will have baffles or something similar to provide more surface area and a longer flow path for vapor to condense. A receiver should be sized based on pump capacity, piping system volume, desired operating pressure range, and desired vacuum storage time.

Condensers and traps. Condensers or traps should be considered if there is potential for a large amount of condensable generation in a given system. As pressure decreases, the rate of evaporation for any given material increases, and this continues until the saturated vapor pressure is reached, where evaporation becomes much more rapid. As long as such material exists anywhere in a vacuum system, the minimum pressure attainable is limited to that saturated vapor pressure.

Furthermore, condensables can cause considerable problems for vacuum pumps. Condensation due to condensable gases entering a liquid-ring pump results in a temperature rise, negatively affecting both the lowest pressure attainable and the capacity of the pump. A high condensable load for oil-sealed and dry pumps, on the other hand, can lead to corrosion and eventual failure. A receiver, as discussed above, can provide a simple gravity and surface-area separation method for handling condensables.

Vacuum traps work on one of two principles: sorption and condensation. Sorption traps use either adsorption or absorption to trap and hold liquid molecules. The effectiveness of sorption traps depends on the interaction between the particular condensables in that system and the trap media, and on the operation of the vacuum system. For example, while alumina balls can efficiently trap oil molecules, any water molecules that pass would be selectively absorbed over the oil and effectively displace any oil molecules that may be trapped. Condensation traps rely on a cold surface to condense vapor molecules. These traps would only be effective if the dewpoint of the vapor is greater than the cooling medium temperature and the pressure is greater than the vapor pressure of the cooling medium. Furthermore, consideration must be taken that the temperature is not low enough to freeze any liquid that may collect on the cold surface.

Pipe sizing methodology

Presented here is a step-by-step approach for sizing the piping in a vacuum system.

  1. Identify and count the number of inlets.
  2. Determine the required flowrate for each inlet. Approximately 1 SCFM can be assumed for laboratory inlets.
  3. Select the location of the supply source, taking into account process piping, venting routes, drainage of condensables, and electrical routing.
  4. Map out the piping layout from each inlet to header to source and approximate the location of all elbows.
  5. Determine the system operating pressure. This varies widely according to the application.
  6. Calculate the equivalent run of pipe by adding friction losses of individual fittings or adding a flat percentage, usually 15–50%.
  7. Approximate line sizes to provide a base value.
  8. Calculate pressure loss using the Darcy-Weisbach equation, given by Equation (6):


Print  (6)




          9. Estimate piping air leakage rates, W , by using Equations (7), (8) or (9), depending on the pressure:


 Print (7)


Print  (8)


Print  (9)


           10. Determine approximate valve locations. Estimate valve air leakage rates, w, by using Equations (10), (11) or (12), depending on the pressure values from Table 1.


Print  (10)


Print  (11)


Print  (12)



   11.  Estimate valve air leakage rates, w, by using Equations (10), (11) or (12), depending on the pressure values from Table 1.

   12. Calculate the velocity through the pipe by using Equation (13):

 Print  (13)



         13. Increase line size when the velocity through a particular run of pipe exceeds 5,000 ft/min, or 4,000 ft/min if noise would be an issue.

         14. Increase line size when pressure drop over entire system exceeds 10% of the pressure at the source.


Other design considerations

Piping design. General design of vacuum system piping closely follows two main variables: pressure drop and condensables. Wherever possible in the design of a vacuum system, pressure drop is minimized. Bends in the system should be kept to a minimum, and all bends that must be installed should be long-radius. To account for any liquid that could get into the system, piping should be slightly sloped toward the receiver tank on both sides of the system. Drainage should be provided at any points in the system where this favorable slope is not possible. Pump exhaust should be directed outside with as little impediment as possible with insulation to guard against condensation and a low-point drain valve. As an additional consideration concerning pumps, two pumps capable of the entire load are recommended for critical systems, whereas one or two pumps designed for a percentage of the load is acceptable for non-critical systems.

As much as generating and maintaining vacuum is an issue, keeping unwanted components out, such as particulate matter and oxygen, can also be a challenge. Sealing of piping systems and pump flanges should be considered to avoid drawing oxygen into a flammable mixture, for example, as well as to minimize pressure drop. Finally, caution should be given to the materials of construction, piping thickness and pressure rating of the system to avoid caved-in lines.

Diversity factor. In the majority of applications, vacuum is not a utility in constant demand. A diversity factor is often used to lower an estimated flowrate to a more reasonable level. However, there is a misconception on how to use a diversity factor that seems to trouble engineers new to vacuum design.

A diversity factor is a ratio of the sum of individual maximum demands of various system subdivisions to the maximum demand of the whole system. Applying a diversity factor to a network of vacuum inlets does not imply that the flow through each inlet on the network drops by a certain percentage. Instead, it is to illustrate that the stated percentage of the inlets have full flow through them at a given time while the other inlets are closed. Applying the diversity factor correctly results in maximum flow through the sections of vacuum pipe where inlets are all open and zero flow through sections where inlets are closed, whereas incorrect application would lead to undersized pipes.

In addition to applying a diversity factor, flowrate and pressure drop can be reduced by eliminating unnecessary valves to reduce air leakage, raising system vacuum pressure, or simply eliminating use-points.

Design for two-pressure delivery. Delivering vacuum at two different pressures poses an interesting problem, and there are a few potential design options (Figure 9). One option would be to design the distribution for the lower pressure, and place regulators downstream of inlets that require a lower level of vacuum. Although such a system could be initially perceived as an efficient method of system design, flowrate downstream would be identical with or without a regulator, as shown in Figure 9.

Figure 9. This drawing illustrates different design options for delivering two distinct pressures in a single system

Figure 9. This drawing illustrates different design options for delivering two distinct pressures in a single system

Alternatively, the system can be designed for the higher pressure, and booster pumps can be added to decrease pressure at inlets that require a higher level of vacuum. This method would allow for lower pipe sizes and a less expensive primary pump; however, it would require purchase and maintenance of additional vacuum pumps. A third option would be to design two separate systems for the two vacuum levels. Two separate systems would require two primary pumps and oftentimes more piping, but it may be worthwhile if the inlets are located at a distance or volume demanded at each of the two levels is substantial.



1. “Vacuum Systems,” Continuing Education from the American Society of Plumbing Engineers, December 2012; Reprinted with permission from “Plumbing Engineering Design Handbook,” Vol. 2, American Society of Plumbing Engineers, 2014.

2. Agilent Technologies, Inc., Agilent Rotary Vane Pumps, DS-102 Rotary Vane Pump Speed Curve, 2015, reproduced with permission.

3. Herring, Dan, “Oil Sealed Rotary Vane Pumps, Part 1,” Vac Aero International Inc., 8 February 8, 2016.

4. Curati, Marino, others, Renewable Power with Rotary Lobe Pumps, Pumps & Systems, April 2013.

5. Walas, Stanley, “Chemical Process Equipment Selection and Design,” Butterworths, 1988.

6. Emerson, “Vacuum Control,” Technical. Vacuum Regulator Installation Examples.

7. Taken from: Coker, A. K., “Ludwig’s Applied Process Design for Chemical and Petrochemical Plants,” 4th ed. Vol. 1, Elsevier, Burlington, Mass., 2007; Original: Ryans J.L. and Croll, S., Selecting Vacuum Systems, Chem. Eng., Vol. 88, No. 25, 1981: 72.


GovoniPatrick Govoni is a process engineer for DPS (959 Concord St, Framingham, MA 01701; Phone: 508-861-3773; Email:, which predominantly deals with design of biotech manufacturing facilities. He has experience generating P&IDs and PFDs, specifying equipment and designing utility systems. Prior to work at DPS, he worked as a manufacturing engineer at DSM and led to over $200,000/yr in cost savings. Govoni holds a B.S.Ch.E. from Rensselaer Polytechnic Institute (RPI) and is an active member of the Education Planning Committee with the International Society of Pharmaceutical Engineering (ISPE).

Related Content

Chemical Engineering publishes FREE eletters that bring our original content to our readers in an easily accessible email format about once a week.
Subscribe Now
Purdue University Saves $400,000 Annually with Local Vacuum Networks
Bag filter Housings/Vessels
Innovative Backwashable Media Filter
Automated Vertical Tower Filter Press
Visit VACUUBRAND at ACS in San Francisco!

View More

Live chat by BoldChat