R. Kleinubing, Emerson, Pompano Beach, Florida
For nearly a decade, digital transformation and the industrial internet of things (IIoT) have been among the most common buzzwords in process manufacturing, particularly in the maintenance and reliability sector. The hydrocarbon industry has long heard rumors of a second industrial revolution, poised to usher in a new era of efficiency and optimization.
However, the reality seems to be something quite different. Many plants are pursuing digital transformation strategies and implementing IIoT solutions sporadically across their assets to improve maintenance and reliability strategies, but few have achieved a true step change in performance. On the contrary, most are still operating in a very similar manner to 10 yr–20 yr ago—maintenance and reliability personnel perform scheduled manual rounds, and analysts then pore over the data to identify trends whenever they have the spare time—albeit with the addition of some digital tools.
If nothing else had changed, maintaining an “If it’s not broken, don’t fix it” policy toward maintenance and reliability would be an effective strategy. However, the hydrocarbon industry has changed dramatically in the last decade, necessitating a shift in maintenance and reliability tactics to stay competitive. Globalization has increased competition for nearly every manufacturer around the globe. Simultaneously, increased public, corporate and governmental pressure in almost every region has driven a need for more sustainable operations that can demonstrate lower energy use, reduced emissions and less waste. All these changes are coming to a head just as nearly every process manufacturing organization faces a global workforce shortage. The most experienced engineers and technicians are retiring in droves, taking their deep expertise of the plant’s assets and operations with them.
To navigate this new world, organizations must find ways to bring in new technologies to support the limited personnel they have, but without increasing the complexity of those people’s jobs or overwhelming them with raw data they are not trained to use. The solution is to implement a predictive maintenance technology plan founded on a boundless automation vision of seamlessly moving contextualized data wherever it is needed. When asset health data from both critical and essential assets can move seamlessly from the intelligent field through the edge and into the cloud, reliability personnel have access to the intuitive, actionable information they need to make better decisions and intervene well before assets fail and processes or plants shut down.
The case for predictive maintenance. As companies face increased competition in a global marketplace, every dollar they spend counts more than ever. As a result, spending is under increased scrutiny, and changes to the status quo are often met with skepticism, especially if they increase costs. Therefore, plant management often considers predictive maintenance technologies as nice but not necessary. If a plant does not have a vibration monitoring system installed on a pump, that pump will still work—though less efficiently—even if it is cavitating or off balance. Because the flaws detected by predictive maintenance systems are unlikely to stop a facility immediately, the technology is often seen as a luxury the plant cannot afford.
Such a position has been possible for a long time because for decades many plants have had a deep bench of experienced personnel who could move around the plant attending to issues and making educated judgments as to when assets would fail. Today, however, this is becoming less possible with each passing year. The technicians with decades of experience necessary to listen to a pump and ascertain if something is wrong by the sound it makes are leaving or have already left. Not only are replacements hard to find, but the new technicians that plants are able to hire typically have far less hands-on experience necessary to diagnose problems without an expert by their side—and those experts are gone.
Moreover, even if a plant still has a wide array of experienced technicians and analysts who can diagnose problems, compliance with regulations has made run-to-failure a dangerous approach. Today, an asset failure can lead to an environmental incident that causes severe financial penalties and extensive reputational damage.
To address these and related issues, today’s forward-thinking organizations are implementing predictive maintenance strategies based on online continuous condition monitoring to eliminate disruptive events and allow them to do more with fewer people. Powerful wireless sensors—many of which have built-in analytics—are small, affordable and easy for plant personnel to install. The cost of the equipment is quickly offset by the fast return on investment (ROI) that stems from avoiding unplanned outages and by eliminating time wasted on low-value tasks, like data collection. To realize a fast ROI, plants should seek out predictive maintenance solutions that provide standardized connectivity and delivery of intuitive, actionable information.
Select for standardized connectivity. One of the key mistakes that can make implementing predictive maintenance more costly and complex is selecting solutions that are not designed specifically for industrial environments. For example, WiFi is a convenient communication technology, but it struggles to maintain a reliable connection in complex process manufacturing settings. The best solutions will instead use industry-standard protocols like WirelessHART, which have been designed specifically for process plant applications.
Selecting sensors that use fit-for-purpose industrial protocols not only ensures more effective connectivity, it also helps reliability teams standardize across the many variables they monitor. Instead of having different protocols for vibration, pressure, flow, temperature and other sensing devices, they have a single standard across all devices, making every variable easy to integrate into the same technology infrastructure. In addition, industry-standard protocols like WirelessHART are designed for cybersecure operation, a critical improvement over sending sensor data to the enterprise cloud using WiFi.
Prioritize timely, actionable information. As more plants have been reduced to significantly smaller crews, the data collected on scheduled rounds has become difficult to use in a timely manner. If a plant has one or two people collecting data on 100 assets via a walkaround, the process can easily take 2 wk. When data collection is finished, analysts typically spend at least another week analyzing the data and returning their findings and recommendations to maintenance crews. Between the initial readings and the final report, the team has had a 3-wk gap during which small problems could have evolved into asset failures.
With modern predictive maintenance technologies, those same technicians receive all the data they need on their workstation or mobile device. Even developing failures too small to be discernable by a technician’s senses will be identified and flagged. In the most advanced systems, asset health will be automatically categorized by severity for easy prioritization by technicians of any experience level. Assets displayed in green are healthy and need no attention, assets in yellow have developing issues that should be explored and assets appearing in red have critical issues that should be addressed immediately.
Teams can use this information to more effectively schedule maintenance around planned outages, helping avoid unscheduled shutdowns of equipment. If the team knows that a 1-hr stoppage is coming while equipment is changed out for a new process, they can more effectively use that time to perform maintenance if they have timely, accurate data, so they know exactly what they need to do and how much time they will need to accomplish it (FIG. 1).
Many plants are also turning to edge analytics devices for increased visibility of the health of essential assets. These sensors apply embedded analytics to all readings to alert maintenance and reliability personnel to the most common faults in their assets, freeing analysts to focus their time on more complex issues.
Predictive maintenance in action. For greenfield projects, planning to incorporate predictive maintenance technologies from the earliest stages can help companies ensure increased reliability from their first day of operation. For one North American company working on a new project, incorporating predictive maintenance into their project helped them identify and eliminate bottlenecks, and do more with the limited expert staff they were able to find in their earliest days.
The project team worked closely with its automation supplier to gather the people responsible for maintenance and reliability, as well as the original equipment manufacturer (OEM) suppliers involved in the project to train everyone on how to select the right equipment and sensors for their assets. Once the solutions were selected, the automation supplier trained users on how to analyze the data and route it most efficiently to the right people.
The company has since expanded with additional projects, but still retains its core diagnostic group in a central corporate location. That team receives data from many different sites worldwide, performs the necessary analysis on the data and reaches out to individual sites with recommendations. Those same sites are also bolstered by data from edge analytics devices, which provide them real-time updates on their onsite equipment with actionable information so local technicians can step in and perform repairs—often on the most common issues—as necessary (FIG. 2).
At the enterprise level, the company cross-references all machine health data with process data from the plants’ distributed control systems to track and trend how machine health is impacting performance and quality.
Brownfield plants also benefit from predictive maintenance technologies. At any established plant, the local technicians can almost certainly identify the bottlenecks and most troublesome assets. What they might not know, however, is how to understand that failure early enough so they can properly plan the repair. These existing systems are just as simple to retrofit with wireless sensors, providing teams the data they need so they are not just avoiding accidents, but also outages.
Prediction drives productivity. As hydrocarbon organizations look to the future of their operations to meet new sustainability benchmarks and secure competitive advantage in a global marketplace, many are seeking ways to make the most of the resources—both personnel and technology—they already have available. Implementing predictive maintenance solutions provides a path to accomplish that goal.
Even a small staff can safely and effectively monitor plant assets—not just the most critical, but essential and balance-of-plant assets, as well—both to ensure they stay operational whenever they are needed and to drive more effective scheduling and eliminate low-value tasks. Personnel will be safer and upskilled faster because they will have real-time visibility of asset health alongside decision support to help them take the most effective actions in any situation. More effective maintenance with a fast ROI is possible for those that implement the right technologies to make it a reality. HP
Romeu Kleinubing serves as Manager of Global Sales Enablement for Emerson’s Process Systems and Solutions. A seasoned global business leader, skilled in integrating technology and people to solve real-world problems, with global multi-industry (including mining, pulp and paper, chemical and petrochemical), multi-language (fluent in Portuguese, Spanish, English and Italian) expertise, Kleinburg has extensive experience in global product strategy and team leadership. He earned a Bch degree in mechanical engineering and an MBA, as well as a vibration condition monitoring (vibration analyst level II) certification.