Compressed Air Systems: Optimizing Reliability – Minimizing Risk

July-August 2000

R. Scot Foss
Plant Air Technology
Charlotte, NC

There is a natural assumption that having decent equipment and taking care of it will guarantee reliability in the system. From this foundation we build on the next error in thinking which is to believe that this effortful approach will net a minimum risk of interruption in the facility. Were it only this simple.

Those of us that have operated compressed air systems for any period of time know that they are one of the most costly and unpredictable of all the utilities in the modern plant. As we head into this look at systems, my best advice would be: If you are not going to take a thorough and systemic approach, buy the least expensive equipment that you can buy. When it fails to meet your expectations, you will not feel nearly as bad as you would have had you purchased remarkably expensive stuff and netted the same result. Over the years, we have found that a well-developed system can tolerate less than spectacular parts and still produce excellent results. Clearly, the best system would be one, which has been thoughtfully configured with the most reliable equipment available.

One of the dilemmas is that the compressed air industry sells features and benefits. It is nearly impossible to determine the shortcomings that you may be faced with based on a discussion with most sales personnel. A basic truth is that it is knowledge of the shortcomings, not the features that will best facilitate an excellent result. You can maneuver around the bumps in the road, when you know that they are there. There isn’t any perfect out there. What usually happens is that the individual responsible for the plant air systems decisions bases them on experience. This is OK as long as you don’t define experience as what you got when you didn’t get what you wanted. We have seen more systems based on bad experience than those based on sound engineering. Often this approach results in each individual item being evaluated on its own merit, rather than its ability to integrate into the solution desired.

Risk Planning

One of the first questions is how will we define the term “risk”. When we poll operators, plant engineers, and maintenance personnel in workshops, the majority tells us that failure in the system is totally unacceptable. Any interruption, curtailment, or negative impact is failure, therefore a lack of reliability. After all, we are talking about rotating equipment. It is not a question of whether it will fail. It is a question of the impact of the failure on productivity. If you develop the system for 100% reliability and 0% risk, you are going to spend a lot more money on the capital equipment, which is no guarantee against failure as it has been defined.

Another issue is that production or process could, but seldom participates in managing the curtailment associated with an element failure in the system. Mature organizations develop plans to limit the least strategically required demand so that the balance of the system can be continued in normal service. Another element of this type of strategy is to have no unit so large that its failure can limit more than an acceptable curtailment. In many industries, the operating protocol is never to operate a compressor or dryer that is so large that its failure could cause an interruption of any kind. For most this means you need at least four or five operating units.

The next step in our failure scenario is to be able to isolate a percentage of the system equivalent to the worst potential failure. This can be accomplished in an orderly and automatic format, so to protect the on-line production or process equipment while this limited portion of the system turns down.

There are a number of systems where the piping is configured so that the various production areas can be isolated with slow acting valves managed through an SLC that prioritizes production. Management can change the priorities on a daily basis if desired.
In a just in time environment, inventory on hand can manage a limited curtailment. The last element of the plan should determine the maximum allowable, time-weighted, curtailments and the allowable duration. One major industrial corporation plans on curtailment not to exceed three times per year, not to exceed 25% curtailment with no more than four hours per incident. This can be planned and managed. We find that a mature and defined approach towards risk management allows the designer and operator of the system to minimize capital and operating expense over more traditional protocols.

A common approach to prevent system’s failure is to operate too much part loaded or throttled power all of the time at as much as 20% pressure than the minimum acceptable value. This is done so that when a unit fails, the balance will load up preventing curtailment. The added pressure will support the loss of supply volume and the minimum operating value during the response. This is done 100% of the time to manage a limited failure event a slight fraction of 1% of the time. More appropriate solutions will be discussed later in the article.

Expectations and Operating Philosophies

Nothing can be more frustrating than the attempt to operate a system where the expectations can rarely be achieved or you operate below minimum acceptable results a portion of the time. This is all too common a situation. You are operating compressors designed for 100 psig with a drying and filtering system with an actual differential pressure of 10 psig. Production expects equal to or more than 90 psig 100% of the time. This is a losing proposition. The only way that you can meet this condition is to install parallel filtration and drying to reduce the differential or operate more capacity part loaded than what is needed. In the modulating mode, this will cause the system’s pressure to rise above 90 psig at full demand. Using compressors to raise the pressure to 95 psig will require twice as much power as necessary, while operating the equipment well off its optimum design. One result will be premature bearing failure of the compressors and freeze up of the refrigerated dryers. The system’s pressure will more widely fluctuate and you will find water condensing downstream of the dryers. Re-rating the compressor to a higher pressure is very expensive and hardly the appropriate solution. The problem is the expectation. The situation is the symptom of the problem. Ninety five percent of the time, we are attempting to correct the symptom without ever defining the underlying cause. Most utilities providers would prefer to operate the supply system beyond its capabilities rather than risk challenging the expectation. Most of the time in auditing the system, we find a $100,000 solution to a $100 problem.

Typical Problems that Reduce Reliability

The following are some of the typical problems with the real underlying root causes:

  1. Systems pressure drops below supply expectations.
    a. A critical point of use application has a dirty line filter.
    b. A regulator at a production point of use has been increased by the production worker. The higher pressure increases the flow to the air user. The higher flow increases the differential across all of the point of use components as a square function of flow. You can increase the regulator pressure and reduce the article pressure at the final point of use. The solution to elevating the article pressure was to change the regulator and other installation components to the next nominal size. This would increase use pressure without increasing the supply.
    c. If the article pressure dropped 10 psig during the use of air and you wanted to elevate the pressure 5 psig, you could double the point of use storage downstream of the regulator. The pressure drop would reduce in half.
    d. You have installed a high rate of flow, short cycle air user in an overloaded
    subheader or branch line. When the new user actuates, a critical pressure user complains about low pressure. You can dedicate storage with a check valve to the critical user. You could also dedicate storage to the high rate of flow user with a metering valve to flatten the rate of flow on the upstream side. This would negate the effect of the new user on the rest of the system. The traditional diagnosis is insufficient supply with the belief that the solution lies in adding more supply capacity.
  2. The pressure dew point fluctuates to levels well above the minimum acceptable results with water condensing downstream in the system.
    a. There is a receiver upstream of the filter and the dryer. The dryer was sized for the compressor/s but didn’t include the surge capacity of the storage tank. The velocity on peak events exceeds the capacity of the dryer. There is not sufficient contact time for drying.
    b. A large transient demand event causes a suction negative rate of change on the downstream side of the clean up equipment causing a high differential and carryover to the system.
    c. One or more drain traps have failed.
    d. The dryer was selected based on theoretical conditions with the coolers clean and the air temperature and relative humidity at the high average, but not the peak. The normal condition is dirty coolers. Although the peak conditions only occur for a few hundred hours per year, it is sufficient to create a reliability problem for production. You must design for your worst ambient and maintenance conditions.
    It would not be unusual to determine that you have the wrong type of dryers leading to the replacement with lower dew point dryers which can cost up to five times more operating cost that the original dryers. Without the problem definition, the original dryer which should produce a 40F dew point produced a 70F dew point. Because complaints of water in the system have ceased, everyone is delighted.
  3. One of your large centrifugal base load compressors frequently surges causing the pressure in the system to shut down critical production applications.
    a. A large volumetric event shuts down from time to time. When it does, the compressor control valves cannot stroke in sync with the event. The result is a surge and shut down of the compressor.
    b. The compressor was specified based on the manufacturer’s conditions rather than the actual range of conditions for the site including water temperature, air inlet temperature, relative humidity, and fouling factors for the coolers, inlet filter, and impeller dirt loading. The natural curve drops during all off design conditions exposing a lower surge pressure on the unit.
    c. A marginal sized dryer and aftercooler are selected on a price sensitive purchase. The mechanical contractor installs the signal line downstream of dryer. The start up service adjusts the controls for the compressor discharge pressure rather than downstream of the dryer. The differential across the dryer and the aftercooler are absorbed upstream of the signal location at the elevated control pressure. The results are a final stage compression pressure which is equivalent to the design pressure of the compressor plus the clean up equipment differential. The control panel on the compressor reads “line pressure” which is actually the signal pressure downstream of the dryer. No one suspects a problem, but the unit surges on a regular basis. You can move the signal, but the best choice is to lower the control pressure to the desired system’s pressure. Make sure that you create a process flow diagram with all of the process values on it including the set points. Show the signal line and make sure that you know what the desired results need to be. Also request “0-0%” performance curves for the compressor based on your metrics and site condition extremes so that you can properly evaluate the best choice of compressor for the application. It is not unusual for the original equipment supplier to suggest that this problem can be solved with a PLC control retrofit or a total automation system ranging in price from $25-$150K. In some situations, adding storage to the system can control the rate of pressure rise in the system to synchronize with the valve stroke speed of the compressor controls. In some cases, the event application in production that causes the problem can be retrofitted with a slow acting shut off valve that eliminates the rapid rise in system’s pressure. “

Good Trouble Shooting Rules

There are dozens of scenarios like the previous examples of missed diagnosis. The following are good rules of thumb for trouble shooting a compressed air system:

a. Whatever the problem appears to be, its probably something else. Air systems always display the symptoms of the problem. You have to dig to find the problem. It can be much like peeling an onion.

b. The first solution to the problem is seldom the best solution. Air systems are very interactive and offer nu-merous approaches to each critical issue. If you don’t have at least three optional approaches to the problem, you haven’t investigated adequately.

c. You have to draw a picture of the problem in order to understand what is actually going on. The dynamics of the system are too complex to grasp between your ears.

d. The only “Bigger is Better” in the air system is storage capacitance. Bigger compressors, dryers, and even piping are not better.

e. Whatever you do in the system, do it slowly, other than responding to demand. Satisfy demand with potential energy and demand controls and then wait as long as possible to replenish the storage. The longer you wait the least amount of energy you will use. When the pressure drops in the system, control the rate of decay so that you can wait as long as possible to add the next available compressor. The longer you wait, the more likely the event that caused the decay or another air user will stop allowing the pressure to recover without the add. Wait, wait, wait!!!

This ability is available through a combination of software, control storage, and improved control set points.

Rules for Reliable & Efficient Systems

The following are a series of excellent rules for the development and operation of compressed air systems:

a. The larger the individual compressor relative to the total demand, the higher the likelihood that the system will fail and cause production interruption. Never select a compressor that is so large that its failure can cause the system to fail.

b. Select your base load compressors for the best displacement to power efficiency. Select your trim compressors for cold start permissive speed to full load and automation flexibility. Remember that the faster the trim compressors, the longer you can wait to add.

c. Make sure that the total trim capacity of the system is equal to or larger than the largest base load compressor. This will assist in allowing a back up base compressor to come up on line in the event of another base failure.

d. Know all of the shortcomings of the equipment you plan on using before buying or applying it.

e. Write a number of detailed failure scenarios including the elements of supply, demand, events, temperature, pressure, capacitance, and time. When you know what can go wrong precisely, you can develop a plan to manage failure without a production interruption. You can’t buy reliability, but you can plan and design for it.

f. Parallel as much compression and treatment equipment as possible so that the failure of any part of the system will not take out an entire compressor and dryer-filter train.

g. Benchmark and trend all variables and deltas against design values. This is the best method of predicting problems in advance of failure.

Know the Right Questions to Ask Before You Buy

Reliability is certainly an important issue. It is important enough to ask the questions of potential suppliers to make the best possible decision regarding what you may be faced with if you select their equipment.

a. How much running maintenance downtime is required per year?

b. Does the manufacturer require annual inspections? Can you train in-house people to do this work? How much out of service time is required?

c. Will anything that can occur as a result of normal wear and tear or fouling cause the performance of the machine to degrade? How does this occur? What kind of performance loss is possible?

d. What other compressor capacities and pressure are available in the same frame as the one that is being presented? We would strongly recommend caution in selecting a unit above mid-frame unless you are keen on maintenance.

e. What tread able information can we measure and what tests can be conducted that will anticipate performance degradation?

f. What is the expected life between major overhauls or repairs? What do most overhauls cost in today’s dollars?

g. What are the ways that this equipment will fail? What is the typical root cause for these failures? Is there any option, which can offset or correct these potential problems? What costs can I expect for these options? Are there any applications or operations advice which will extend the potential for these somewhat predicable failures? How much lead time is required for parts needed and repair work that needs to be done?

h. Does the local service dealer maintain an inventory of parts for the potential for both running maintenance and first level failure repair?

i. If you have to change any major components or perform a major overhaul, what are the risks associated with loss of original performance? If you change out a major stage part such as an airend, do you have to change out the other stage? Are they matched? Will the unit need to be tested before it is restarted after the repair? What test protocol will best determine the performance?
There isn’t any perfect out there. Get familiar with the downside of the equipment. Tell the sales representative that you want the good and the bad. It isn’t an unreasonable request. It will help you with instrumentation and maintenance planning.


Whether you are designing a new system or retrofitting an existing compressed air system, you can control the reliability, and manage limited curtailment at the precise level you desire. You can control the front end, maintenance, and operating costs of the system. Remember that all undefined problems are not insufficient supply. This does not have to be dealt with as an art form. It is a science. With the knowledge of good system’s technology coupled with the up and downside of the components applied, compressed air can contribute in a positive way to improved productivity at lower costs.

July-August 2000