Maintaining compressed air systems

Predictive maintenance and controls in the information age

Unanticipated compressor outages can be one of the most frustrating things to happen in the plant. Similar to electrical power failures, compressor outages disrupt production, require extensive repairs, and lead to the many associated costs of unscheduled downtime. However, many of us know that these outages can be identified and avoided by using statistical trends based on daily maintenance data readings.

Unfortunately, as many companies expand operations faster than they expand their maintenance staff, time is pressed and daily data collection becomes less of a priority compared to other maintenance duties. Additionly, the practice of having back-up compressors is more of the exception than the rule. This has led more and more plant operators to seek alternatives to the time-consuming, but necessary, practice of data collection and monitoring. This article reviews where predictive maintenance programs originated and present options that companies may utilize to streamline the data collection process.

The importance of daily data collection
Generally, the operation of a compressor, like other equipment, is constant and usually predictable. Compressor applications and subsequent performance are based on specific physical conditions. Over time, the physical and mechanical demands of operation adversely affect the general performance efficiencies of the compression process and operating temperatures.

For example, a condensate trap on a two-stage unit that fails allows the intercooler to fill with liquid. Consequently, the velocity of the compressed air sweeps any foreign liquid into the next
compression chamber. These result in premature compressor wear and air system contamination.

To avoid this and other similar scenarios, most companies typically perform periodic observation maintenance programs for compressors and other rotating equipment. These programs require that readings of temperatures, pressures, and functions be recorded. Then, by analyzing this data, operators can schedule downtime to address maintenance issues.

Control systems
Technology has played a significant role in improving predictive maintenance practices. Let's begin with the first step in the process, the data collection source. Typically, operators used to be required to manually record all gauge readings from the compressor on a daily basis. For instance, compressors that have electro-pneumatic systems gauges monitor several functions on the compressor. These gauges sometimes are unlabeled and occasionally require operators to gather multiple readings to ascertain compressor functions. Now, electronic, or microprocessor controls, offer detailed text information on compressor functions on one control panel. By incorporating the controls operators obtain actual operating values. Often operators perform this function from one central location at the touch of a button.

In addition, the information available from the microprocessor controls is more accurate. Instead of common pressure and metering devices, the microprocessor control system relies on electrical transducers and sensors. These devices sense air pressure and temperature values, which are then transmitted to a central microprocessor.

In turn, the microprocessor interprets the information and adjusts the compressor's output through an integrated control system. The microprocessor also measures and stores compressor-operating data for future maintenance reports and needs.

For instance, monitoring the airend discharge temperature of rotary screw compressors can be a critical element in reducing downtime. In this case, the microprocessor control system alerts operators to any changes in these values to allow for preventive maintenance.

The microprocessor control system also allows operators to adjust shutdown setpoints automatically and respond to alerts. The task of mechanically resetting each protective switch is no longer necessary. The microprocessor allows resetting from the control panel.

When compressor units experience shutdowns, it is sometimes difficult to pinpoint the root cause because several alarms may have been activated. However, microprocessor control systems can monitor multiple alarms, and if shutdowns occur, they can recall the alarms to help identify the various problems and points of origin.

Further, the microprocessor control system provides troubleshooting assistance through its monitoring alarm system. Even if a warning alarm is activated when the compressor is unattended and the system corrects itself, the alarm remains on the microprocessor panel along with the various operating parameters that were present at the moment of the alarm.

This enables operators to reconstruct and evaluate the conditions when
the warning alarm was activated. Effectively, predictive maintenance technology is or can be built in to the equipment by manufacturers.

Data logs
Once the data is collected, the next step is logging and trending the information to plan preventive or corrective maintenance practices. Historically, companies typically performed these daily data collection and routine maintenance programs for compressors and recorded them in sequential
log sheets.

Unfortunately, log sheets provide only a limited area for recording system performance. When filled, the log sheets usually are filed and only reviewed after an unscheduled failure occurs.

Currently however, log sheet data no longer needs to be collected and viewed from a historical perspective. Operators or maintenance staff can load the data to computer spreadsheets and trend analysis programs that offers operators the option of viewing visual charts.The charts allow tracking compressor performance and help identify needs for servicing.

However, even with microprocessor control systems, log sheets still remain a key component in preventive maintenance. To determine sources of problems, log sheets can be examined with computer trend analysis data, such as intercooling functions within the compressor process.

Intercooling is critical to both a centrifugal compressor's performance and the life of internal parts. Larger compressors use water-cooled heat exchangers to achieve efficient heat transfer. Often, minerals and solids suspended in the cooling water collect in the cooler and reduce the heat transfer capability and efficiency of the compressor.

While a trend of increasing temperatures may not be noticeable on the log sheet, computer-generated analysis and graphics identify the need to revise cleaning and back flushing schedules. Analyzing maintenance observations and statistical data, supported by trend graphics, enables plant operators to relegate unscheduled system outages to the routine maintenance program.

Another example of the benefits gained through computer-based trend analysis involves the universal measurement of cooler performance. Commonly known as cold temperature difference, engineers determine this measurement by calculating the inlet temperature of the cooling media and the discharge temperature of the air at each cooler.

The temperature measurements that are required to calculate cold temperature differences are routinely noted in log sheet records, but the calculations are often postponed or overlooked until a problem occurs. However if operators plot the two temperatures, the data quickly reveals the trend in cooler performance that is useful for future planning purposes.

Generally, log sheet data and observations for any rotating piece of equipment can be classified as qualitative and quantitative. The qualitative observations are quite simple.

For example, either a condensate trap operates or it doesn't. Yet, quantitative observations that illustrate trends used for planning future service and general maintenance schedules are sometimes more difficult to see.

These days, savvy plant operators harness the power of computer programs to identify and analyze these quantitative observations. The combination of daily log sheet entries and compressor control data provides operators with sound and predictable maintenance programs.

Remote access and outsourcing services
There is a new breed of services available to the plant and asset care manager that takes the computerized performance observations to the next level by adding remote access. Communications protocols, such as MODBUS, allow many facilities to download
data onto analysis tools to help predict maintenance schedules.

Compressor original equipment manufacturers and other third-party vendors are working to expand the parameters of microprocessor controls, multiple compressor control systems and predictive maintenance data collection tools to allow plant operators to monitor compressor functions remotely, adjust settings, and collect and trend data. This process allows operators to use one or two tools to monitor the complete compressed air system performance and predict its required maintenance. Shrewd observers will notice that this concept is consistent with Deming's thoughts regarding increased productivity through elevation of the level of technology.

Beyond monitoring the compressed air system, companies also have the option of tying the compressed air system control systems into facility-wide monitoring systems that allow for trending and remote access. Predictive maintenance and control system vendors will be able to evaluate a company's facility and maintenance needs, collect the appropriate technical data, and develop the communications protocol that ties in all the systems into one data collection and trending device.

While the system can be extremely efficient in terms of monitoring equipment, it can be costly to develop. Also, this requires the expertise of a supplier who is intimately familiar with the nature of the equipment in the plant. The demands of the language or protocol are such that they must allow the computers to speak to the compressor's possesive control systems and this may require some custom programming before any predictive maintenance can occur.

The other concern with on-board predictive maintenance technology and custom software packages is that suppliers need to understand the compressed air equipment design in sufficient depth to determine the proper equipment set points to make the predictive maintenance system effective.

The other option that will soon be available for companies is the ability to completely outsource their compressed air system predictive maintenance programs. Similar to outsourcing janitorial services or other maintenance functions, companies will have the option of outsourcing daily data collection, trending, troubleshooting, routine maintenance and scheduled repairs to a vendor who collects the data daily through a modem line.

By utilizing the advancements that have been made to microprocessor controls, industry leaders are working to develop programs that supply the communications hardware to transmit the compressor data to an outside service vendor.

The service vendor not only will collect and trend the data, but also will handle the routine maintenance tasks from filter changeouts to more extensive maintenance needs, including cooler cleaning.

Ideally, this service would be offered as an incentive for predictive maintenance packages, eliminating the need to incorporate compressor control systems into facility-wide controllers and trending tools. As technology improves so do the options companies have for automating these routine but critical predictive maintenance processes. Who knows, maybe technology will prompt the invention of a self-fixing air compressor.