An industry with aging infrastructure and increasing demands needs better monitoring technology.

Between 2009 and 2012, the U.S. faced more than 1,700 refinery shutdowns, an average of over 1.16 shutdowns per day, and almost "92 percent of the maintenance-related shutdowns were unplanned."1 So a proper maintenance strategy is critical. The goals for any maintenance strategy are to:

  • boost revenue through increased uptime and asset maintenance optimization
  • reduce costs by decreasing warranty repair expenses, extending mean time between failures, reducing the frequency of unscheduled downtime and optimizing the workforce
  • manage risk and increase safety by minimizing worker contact with dangerous machines and environments and preventing catastrophic failures
Image 1. Between 2009 and 2012, the U.S. faced more than 1,700 refinery shutdowns, which is an average of more than 1.16 shutdowns per day. (Courtesy of National Instruments)Image 1. Between 2009 and 2012, the U.S. faced more than 1,700 refinery shutdowns, which is an average of more than 1.16 shutdowns per day. (Courtesy of National Instruments)

Increasing production while decreasing the total number of refineries in the U.S. makes these goals even more important.2 Creating a maintenance strategy with the appropriate investment in the right maintenance techniques is not easy. Considerations vary widely but can include inconsistent diagnostics, lack of insight, aging plants, postponement of new equipment, aging workforce, safety, and time or money spent manually collecting data rather than analyzing it. Against this backdrop, maintenance managers are increasingly focused on reducing costs and optimizing efficiency, equipment utilization and worker safety without spending more money.

The sheer volume of refinery shutdowns should be enough to spark a change in the traditional maintenance mix, but other factors are in play as well. Refinery age has become a legitimate factor. The average age of the 15 newest refineries in the U.S. is more than 31 years old, with the newest 100,000 barrel-per-day producing refinery being 37 years old.3 This advanced age, among other factors, can lead to unplanned or uncontrolled shutdowns, which cost the company millions in lost production and potentially millions in fines.

Further increasing the complexity of the issue, the reasons for shutdowns are not always the same. Maintenance causes 23 percent of shutdowns, mechanical issues cause 46 percent and electrical problems lead to 19 percent (more than 60 percent of electrical were from unspecified causes).1 Traditional measurements are no longer enough to account for all of these types of shutdowns; electrical measurements are also necessary.

Measurements such as power, power quality, transformer oil temperature and motor current signature analysis on electric motors can give greater insight into the health of the refinery. Traditional maintenance methods are not sufficient to deal with the volume and myriad of measurements and measurement types required to increase efficiency and prevent unplanned shutdowns.

Collecting more data and information about machine health is only part of the answer. A recent article by International Data Corporation indicates that out of all the data collected, only 22 percent is documented well enough to be analyzed. Of this data, only 5 percent is actually analyzed.4 It is no wonder that 78 percent of maintenance managers are unhappy with their current maintenance approach, according to a survey conducted by Allied Reliability Group. Too much time and money are spent with a lack of results to show for it.

Another challenge that arises is how to scale a maintenance strategy to incorporate all vital assets. Hiring more people is not always a fix. According to the Bureau of Labor Statistics, growth in vibration specialists will only be 4 percent over the next few years. So finding the right people will be a challenge.

Spending more money on traditional automated measurement systems will not work either, because those systems are too expensive and inflexible to measure several different types of signals.

But according to Bernie Cook, director of maintenance and diagnostics at Duke Energy, "Without additional resources or new equipment condition monitoring technologies, the risk of equipment failures will increase." In other words, plants must find a better way to detect problems earlier, because it can mean the difference between a $5,000 bearing replacement and a $150,000 catastrophic bearing failure.

Now is the time to invest in a new strategy. Facilities need a predictive maintenance strategy that integrates with their existing enterprise infrastructure and automates the collection of data on more machines to predict machine failure in advance of catastrophic and costly repairs. A study by researchers at the Electric Power Research Institute found that a predictive maintenance strategy costs about one-third of a preventive or scheduled maintenance strategy and all but eliminates the possibility of catastrophic failure.5 This evidence is buoyed by findings of researchers at McKinsey & Company: "When companies inject data and analytics deep into their operations, they can deliver productivity and profit gains that are [5] to [6] percent higher than those of the competition."6

Oil refineries often require both traditional measurements and power measurements. Refineries need a platform that is flexible enough to incorporate different measurements yet economical enough to produce a return on investment. Some platforms can even be flexible enough to perform both control and monitoring to further simplify systems.

Some products provide a platform-based, ready-to-run integrated hardware and software solution for online condition monitoring that helps companies gain insight and improve operations and maintenance programs by understanding the health of their machinery and equipment.

This extensible suite acquires and analyzes measurements, generates alarms, provides the ability to visualize and manage data and results, and simplifies remote management for large numbers of monitoring systems. It offers:

  • built-in analytics and alarming capabilities and remote interactive visualization of waveform data
  • scalability to hundreds of acquisition devices per server
  • support for vibration and process measurements using integrated hardware options
  • remote management of monitoring systems and users
  • IT-friendly security features that include user and device authentication


  4. EMC Digital Universe with Research & Analysis by IDC—The Digital Universe of Opportunities: Rich Data and the Increasing Value of the Internet of Things
  5. The Electric Power Research Institute has calculated comparative maintenance costs in US dollars per horsepower (hp) for each maintenance strategy.

4 Common Maintenance Approaches

  1. Reactive. Using this approach, facility employees wait until something breaks to fix it. Although the upfront costs for this approach are low, hidden costs such as safety, unplanned interruption and ripple effects to other machines in the critical path can be high.
  2. Scheduled/preventive. This has become the default strategy in many industries, and for good reason. Most professionals think this is the best approach because it is popular and recommended by many machine providers to ensure the integrity of the machine. The maintenance intervals are clear and defined, but facilities risk fixing a machine that is not broken and stopping production more often than necessary. This approach also runs the risk of excessive tinkering that could damage the machine being maintained while increasing operational costs.
  3. Preventive with manual/route-based measurements. This involves periodically measuring key performance indicators such as vibration, temperature, proximity and power. This technique is a good first step to achieving maintenance objectives, but it has its shortcomings. Human error, diagnosing skill, disparate systems and the inability to scale prevent this from being a great solution.
  4. Automated or online automated diagnostics. Smart equipment on machines can help predict issues before they wreak havoc on your budget, safety, efficiency or yield. In an ideal world, an automated solution would be the optimal maintenance strategy. But, as often is the case, it is too expensive and does not have the promise of return on investment for all but the most critical of assets. However, that notion has begun to change.