Get the latest updates on the Coronavirus impact on engineers.Click Here
Compressed Air

Compressed air audit optimizes efficiency

A compressed air system evaluation can provide owners and facilities engineers with the direction they need to ensure their plant is operating efficiently and reliably

By Blake Finch and Luke Streit March 30, 2020
Courtesy: IMEG Corp.

Learning objectives

  • Understand the criteria for evaluating the efficiency of a compressed air system.
  • Identify how air quality affects a compressed air system’s energy use.
  • Describe the different opportunities for saving energy in a compressed air system.

The requirements of an effective industrial compressed air system vary depending on facility size, air quality requirements, minimum demand pressure and other site-specific criteria. When operating or modifying these systems, facilities engineering staff have many different questions to consider, including:

  • Should I make low-cost, minimum-quality air in bulk, then filter/dry it to higher qualities only where necessary?
  • Am I staging multiple compressor systems in the most energy-efficient manner?
  • As demand increases in a facility, how do I know when my piping is no longer big enough?

System evaluations can answer these questions and provide owners and operators with the direction they need to ensure their plant is operating efficiently and reliably. Because each application is different, however, the correct answer for one site might not be the right answer for another site.

Types of compressed air

Compressed air requirements fall into two main categories: air quality and air pressure. Higher-quality air requires better filtration and treatment, which both equipment costs and energy costs, as air must be pushed through these additional components. Higher-pressure air requires higher energy costs as the compressor must work harder to pressurize every cubic feet per minute of air.

A typical evaluation might estimate the efficiency of the system in terms of kilowatts of power required per 100 cfm of air delivered.

A major way to lower the kilowatts per 100 cfm is to lower the pressure drop in the system. System pressure is determined by the required pressure at the most critical process. The minimum pressure required at a piece of equipment is typically outside of your control; it is simply a process requirement. The compressors must generate air at that critical pressure plus whatever pressure losses must be overcome between the compressor outlet and that process. Lowering the pressure drop allows you to lower the compressor discharge pressure.

If you have several pounds per square inch of pressure drop in your piping system due to undersized piping, you are wasting energy. Or, if you are pushing air through unnecessary filtration and other components, you may be forced to set your discharge pressure unnecessarily high. Lowering the compressor discharge pressure by 10 psi will save approximately 5% in energy consumption.

Piping pressure loss evaluation is straightforward; pressure drop can be measured from point A to point B, which allows you to determine if the drop is excessive and how much it could be lowered by increasing the pipe size. A contractor can provide a price for upsizing that section of piping; the piping replacement will either be worth the cost or it won’t, based on the expected return on performance.

Determining how to minimize filtration losses is less straightforward. To do this, you need to know what air quality is required at the various processes in your facility. Many compressed air users in a factory, such as pneumatic tools and some forms of conveyance, do not have stringent air quality requirements. Other applications — such as instrument air, food processing or testing — require very high-quality air.

Compressed air quality levels are defined by ISO 8573-1. This standard specifies purity classes of compressed air in three categories: particles, water and oil (see Table 1).

Table 1: Compressed air quality levels are defined by ISO 8573-1:2010. Courtesy: IMEG Corp.

Table 1: Compressed air quality levels are defined by ISO 8573-1:2010. Courtesy: IMEG Corp.

Particulate classes can be achieved by filtration. Higher-quality filters typically have a higher pressure drop. To push the air through extra filtration, the discharge pressure of the compressor must increase, requiring higher energy cost per cfm.

Oil classes also can be achieved by filtering out the oil that enters the compressed airstream in an oil-flooded compressor. Oil filters require added pressure, which uses additional energy. Oil-free compressors include oil in the machine, but it never contacts the compressed airstream. An oil-less compressor has no oil in the machine.

In either case, because no oil enters the airstream, it does not need to be filtered to achieve the highest class. If oil content is a critical element in your facility, starting with a potentially more expensive compressor that doesn’t put oil into the airstream will save the energy costs of filtering it out later.

Handling water vapor

Water vapor is different from particulate or oil because it is not filtered out. It must be condensed out or absorbed from the airstream — this can be completed by a refrigerated air dryer or a desiccant air dryer.

Refrigerated dryers remove moisture by cooling the compressed air below a certain dewpoint. As the air cools, the moisture condenses out. The amount of moisture removed is directly related to how cold the air can get. This highlights a practical limitation of refrigerated air dryers: they can only achieve dewpoints approaching the freezing point of water, i.e., greater than 32°F. If a refrigerated dryer cooled below 32°F, the water vapor would condense to liquid, then freeze solid within the dryer and piping. To achieve ISO class 3 or better dewpoints, a desiccant dryer must be used.

Desiccant is a substance that has a high affinity for water vapor. (Silica gel packets such as those found in a shoe box or packaged with electronics are an example of a solid desiccant.) The desiccant will absorb water vapor from the surrounding air (lowering the humidity) until the vapor pressure of the desiccant and the surrounding air are in equilibrium. In sealed packaging, a desiccant package can maintain a level of dryness for the product as long as the packaging is closed, because new moisture is not being added to the product.

In a compressed air system, however, a constant flow of humid air is flowing through the dryer. The desiccant will eventually become saturated with moisture to the point that it can no longer absorb water vapor from the incoming airstream. The desiccant must then be replaced or regenerated by a drying process.

When evaluating a desiccant dryer, also consider that the regeneration of the desiccant can consume significant energy. Regeneration occurs by heating and/or purging air through the desiccant to dry it out. Consider the power consumption of an electrically heated dryer when evaluating the total system kilowatts per 100 cfm, as well as the source of “purge” air during regeneration. Most desiccant dryers use compressed air to purge the dryer and these purge losses can be a source of significant energy consumption. They do not affect the energy efficiency in terms of kilowatts per 100 cfm, like other components. Rather, they directly increase the total cfm of the system. Depending on the application, purge losses might cause the most significant energy consumption in a system.

The correct design or remedy for a compressed air system cannot be determined without facility-specific data — and having a predetermined solution in mind may blind you to other possibilities. An audit or evaluation of an existing system is the only comprehensive and reliable way to identify valuable energy-saving potential. The accompanying case studies exemplify the different ways a compressed air evaluation might unfold at different facilities.

Engine test cell facility

This example facility’s industrial compressed air system had an average weekday air consumption of 800 to 900 cfm with peak airflows up to 1,200 cfm. The main compressed air system included 225 horsepower of constant speed compressors and a 100-horsepower variable speed compressor. All compressors were oil-flooded rotary screw type. The air was dried by refrigerated dryers to a dewpoint of approximately 40°F. Filtration for oil and particulates was not consistently applied at outlets of all compressors, so overall air quality for particulates and oil was not accurately known.

This site used compressed air in engine test cells, requiring high-quality compressed air meeting ISO class 2/2/1, which has less than 100 1- to 5-micron particles per square meter, a maximum dewpoint of -40°F and a maximum oil content of 0.01 milligrams per square meter. This high air quality requirement steered the direction of the analysis.

Particulate and oil removal: Particulate and oil filters located at the discharge of compressors met the general air quality requirements for nontest cell applications. Additional point-of-use filters were installed at each test cell to further filter to the air quality required for particles and oil.

Water removal: Water content for general applications was removed by the refrigerated air dryers near the compressors. This style of dryer could not provide the low dewpoint required (-40°F) for the test cells, so a desiccant dryer must be used to achieve ISO class 2 dewpoint of -40°F.

Small, point-of-use desiccant dryers were installed at each test cell in line with the additional particle and oil filters. The desiccant for these remote dryers was regenerated by blowing a percentage of the compressed air through the desiccant and exhausting it out of the filter. A typical point-of-use desiccant dryer might purge 25% of its rated airflow to regenerate the desiccant. These point-of-use filters and driers were points of maintenance, high pressure drop and high purge air losses.

This evaluation sought to determine if higher quality filtration and drying at the source of compressed air would be better than point-of-use filtration and dryers at the test cells. To do this, the greater of the following had to be determined:

  • The added cost of filtering and drying high-quality air to the entire site, including processes which do not require high-quality air.
  • The potential savings of consolidating all the remote filters and dryers to a single source location.

The plant compressors produced air at an estimated average of 18 kilowatts/100 cfm plus dryer power consumption of 0.7 kilowatts/100 cfm.

The remote filters and dryers constituted an approximately 8 psi pressure drop. Removing them could allow for lower supply pressure and a savings up to 4% on the compressor-specific power. This equates to roughly 0.7 kilowatts/100 cfm.

A more significant savings could be realized by targeting the purge losses associated with the desiccant dryers. This wouldn’t change the specific power in terms of kilowatts/100 cfm, rather, it would lower the power consumption by reducing the system cubic feet per minute.

The dryers at the test cells purged approximately 25% of their nameplate rating. Because test cells were the major compressed air consumers on-site, the purge losses at each test cell constituted approximately 15% of the entire system’s compressed air usage.

For this site, investing in new oil-free screw compressors with integrated heat-of-compression desiccant dryers provided significant benefits. The integrated desiccant dryers do not purge any air because they regenerate the desiccant using “free” heat of compression.

Providing this air quality directly at the source meant providing better, higher-cost compressed air for applications such as pneumatic tools. But, providing oil free air at a low dewpoint directly from the central plant allowed for removal of all downstream oil filtration and the subsequent maintenance and pressure loss.

The recommendations of this evaluation were implemented on-site and the specific power consumption of the compressed air plant dropped by approximately 2 kilowatts/100 cfm. Furthermore, the elimination of the purge losses resulted in an air consumption decrease of approximately 180 cfm.

The total average energy consumption for the compressed air system dropped from approximately 170 kilowatts to less than 120 kilowatts, a savings of approximately 30%.

Manufacturing facility with many unknowns

In a second example, a more than 2-million-square-foot manufacturing facility had an average airflow of approximately 2,500 cfm. The main compressed air system included two 250 horsepower constant-speed compressors and a 300-horsepower variable-speed compressor. All compressors were oil-flooded rotary screw type. The air was dried by refrigerated dryers to a dewpoint of approximately 40°F. The compressed air system provided air quality of ISO 8573-1 class 2 or 3 for particulate, class 4 or class 5 for moisture and class 2 for oil. This air quality was sufficient for the known users of compressed air on the site.

An evaluation of the central system was conducted with the goals of identifying any general deficiencies, improving energy efficiency and validating the current compressor staging methodology.

To answer the key questions of this evaluation, data loggers were used to collect information and gather trends over a two-week period. This included:

  • An existing turbine flow meter (which provided the total airflow of the system).
  • Temporary data loggers at each compressor to monitor and record power consumption at 30-second intervals.
  • Pressure monitors at three locations in the compressed air system throughout the building. Pressure at these locations was recorded at intervals to match the power consumption data collected.

This data allowed the engineering team to proceed with an analysis of the existing system to identify any deficiencies.

Evaluation of existing distribution system

As shown in the engine test cell facility, the compressor pressure setpoint should be based on maintaining minimum inlet pressure required at critical process. High pressure drop between source and critical process will require increasing the pressure setpoint, which costs energy. The pressure sensors installed in this facility allowed us to evaluate the pressure drop at remote areas of the factory.

An example of data, measured in pounds per square inch, is shown in Figure 1. Location A was in the main compressed air header near the compressors. Locations B and C were in remote areas at opposite ends of the factory.

Figure 1: Data gathered on the pressure drop throughout the facility in the manufacturing facility case study showed pressure swings in the system as the load changed, particularly during first-shift manufacturing hours. Courtesy: IMEG Corp.

Figure 1: Data gathered on the pressure drop throughout the facility in the manufacturing facility case study showed pressure swings in the system as the load changed, particularly during first-shift manufacturing hours. Courtesy: IMEG Corp.

The collected data showed a pressure loss in the piping system out to the ends of the factory, as expected — an approximate 1 psi drop to location B and an approximate 2 psi drop to the farther location C.

The data showed pressure swings in the system as the load changed, particularly during first-shift manufacturing hours. Interestingly, the pressure drop to remote areas — i.e., the difference between A and B or A and C — did not vary greatly throughout the production day. The swings in pressure were relatively uniform across the entire system.

This indicated that the primary consumer of compressed air causing the pressure fluctuations was near the main trunk line, thus impacting the pressure of the entire system, not simply a remote area. This also indicated that pressure losses in main piping to remote areas was not a limiting factor in capacity and increasing the main piping size would not significantly improve performance; less than a 1% efficiency gain was anticipated.

Compressor power consumption

Power consumption was monitored during the study of the three main air compressors. The original intent of this study was to ensure that the staging of compressors was operating as efficiently as possible. A sample of the data, measured in kilowatts, is shown in Figure 2. In this analysis, compressor 1 was variable speed and higher output than compressors 2 and 3, which were fixed speed.

Figure 2: Data gathered on compressor power consumption at the facility in the manufacturing facility case study showed that compressor 1 was operating as a trim compressor — as intended — and that one or both fixed-speed compressors were operating as lead compressor(s) depending on load. Courtesy: IMEG Corp.

Figure 2: Data gathered on compressor power consumption at the facility in the manufacturing facility case study showed that compressor 1 was operating as a trim compressor — as intended — and that one or both fixed-speed compressors were operating as lead compressor(s) depending on load. Courtesy: IMEG Corp.

The data showed that compressor 1 was operating as a trim compressor — as intended — and that one or both fixed-speed compressors were operating as lead compressor(s) depending on load.

The blue line shows the range of modulation of compressor 1. The flat line areas at the top of its range indicate where it could not keep up with demand. The pressure in the system header would have started to decrease during these times. Fittingly, this flat area precedes the enabling of the second fixed-speed compressor. Likewise, the flat line areas at the bottom range of modulation for the variable-speed compressor show where it could not turn down further. The system pressure would have risen to the point where compressor 2 turned off.

Before going offline, there is also a short stop in the power consumption of compressor 2. This represents where a fixed speed compressor goes from “loaded” to “unloaded” operation. This unloaded operation period still consumes energy without producing compressed air. This highlights the value of a variable-speed compressor in the trim position rather than simply loading and unloading a fixed speed compressor.

The data collected showed that the unloaded duration for compressor 2 before cycling off was short and that significant energy was not being wasted by inefficient cycling of fixed-speed compressors.

Is there any energy-saving potential?

Up to this point, the study had not shown many expected avenues for efficiency improvement. Distribution losses were reasonable, compressors were staging efficiently and no changes were needed to filtration or drying.

Figure 3: Shown is a compressed air header with pressure sensors. Courtesy: IMEG Corp.

Figure 3: Shown is a compressed air header with pressure sensors. Courtesy: IMEG Corp.

An unexpected finding revealed that a minimum of two compressors were always running during the period monitored. Most compressed air usage should be tied to manufacturing processes; we expected to see a significant increase in compressed air consumption during first shift, marginal consumption during second shift, then minimal consumption during periods of no manufacturing activities. However, when we compared the power consumption data with compressed airflow data, it showed significant compressed air usage even outside of normal manufacturing hours (see Figure 4).

We did not expect the airflow to drop to zero because there will always be some compressed air consumption due to leakage in the piping and some manufacturing processes may continue to use compressed air during overnight operation. But the data showed a level of compressed air consumption during off-shift hours that was much higher than should reasonably be attributed to leakage. Approximately 75% of the compressed air costs for the facility did not appear to be attributed to manufacturing processes.

Figure 4: Comparing the power consumption data with compressed airflow data at the facility in the manufacturing facility example showed significant compressed air usage even outside of normal manufacturing hours. Courtesy: IMEG Corp.

Figure 4: Comparing the power consumption data with compressed airflow data at the facility in the manufacturing facility example showed significant compressed air usage even outside of normal manufacturing hours. Courtesy: IMEG Corp.

The engineering team then recommended finding where the compressed air was really going. If this facility could reduce the nonproduction related airflow by even 50%, it would save more than $100,000 annually in energy costs.

Improving efficiency of compressed air systems may save you a small percentage of energy cost per cfm. However, eliminating unnecessary compressed air consumption saves 100% of the energy cost for the cubic feet per minute saved.

Figure 5: An air compressor controller display screen shows discharge pressure trend. Courtesy: IMEG Corp.

Figure 5: An air compressor controller display screen shows discharge pressure trend. Courtesy: IMEG Corp.

Use compressed air? If so, take these steps

This evaluation highlights the importance of monitoring compressed air usage, a task that facility engineers can initiate by asking department managers the following questions:

  • Does your department use compressed air?
  • How do you regulate your compressed air usage?
  • Are there any processes that are enabled via manual valves that could be left open? (A single ¾-inch ball valve left open will leak approximately 1,000 cfm of compressed air at 100 psi.)

If you suspect your system has excessive leakage, walk through the facility when manufacturing processes are off and listen — compressed air leaks will make noise and major ones may be audible during a walkthrough. You can also use an ultrasonic leak detector to find the sources of the leaks.


Blake Finch and Luke Streit
Author Bio: Blake Finch is a project manager and mechanical engineer at IMEG Corp.; Luke Streit is a project manager and mechanical engineer at IMEG Corp.