Remote Monitoring for Asset Reliability

Summary

At the end of the COVID-19 pandemic, many process manufacturers learned an unexpected lesson: under the right conditions, targeted remote work can be significantly more efficient than anticipated. As plants implemented remote systems, they saw that they were able to consolidate resources, empowering a smaller group of people to monitor and maintain assets, processes, and software over a wider global area, all without increasing costs due to travel.  

Perhaps even more importantly, as the pandemic waned and fewer people than expected returned to the process manufacturing sector due to retirements and personnel shortages, organizations learned that in many cases, remote work freed their limited staff to focus on high value tasks. 

A result of this change has been a shift in the way many organizations handle machine reliability analysis. More than ever before, plants are turning to remote analysis programs to supplement or replace onsite machinery health monitoring. For some plants, remote analysis is a more cost-effective way to continue the machinery health monitoring they have long relied upon. For others, it is a doorway to improved operation that would be otherwise out of reach. In either case, understanding the value of remote analysis and following four key strategies to implementation are critical to getting the most out of any remote analysis program. 

The value of remote analysis 

Historically, major assets in a plant—those most essential to operation or those with the potential to create safety hazards upon failure—would be equipped with an online monitoring system. And plants that had enough equipment to warrant multiple systems often employed a few analysts on their reliability team.  

However, balance of plant assets—those that are important to efficient operation but not critical—still require monitoring. Highly efficient reliability teams recognize that just because an asset is less critical, it does not mean it can be allowed to run to failure. The earlier technicians detect problems with an asset, the lower the maintenance cost to fix it. For example, if a team sees a motor bearing beginning to fail and doesn’t act—or worse, is not aware of the flaw in the first place—they risk the entire motor failing, a coupling breaking, or a fire.  

To avoid these types of problems, balance of plant assets were typically monitored through manual rounds, a time-consuming process where an individual would visit each asset with a handheld device and record vibration readings, then return to the office and analyze the data. Teams would set alarms to notify them of problems detected in the data collected from rounds. However, this system was not efficient. In many cases, by the time a handheld device detected a problem severe enough to trigger an alarm, significant damage was already done. The solution was reactive, not predictive. 

To complicate things further, today’s plants operate in a highly competitive environment where peak efficiency is critical to profitability, safety and sustainability. As a result, organizations expect their plant staff to collect data from balance of plant assets, analyze it quickly and solve problems before they interrupt production, all with fewer people. 

Fortunately, wireless technology enables reliability teams to cost effectively implement as many vibration measuring points as they need, without expensive hard wiring, and empowers them to collect the data continuously without the need to commit valuable personnel to walkarounds. Due to this paradigm shift in reliability monitoring, today, more data is collected in plants than ever before.  

But this new normal creates its own set of problems. Massive amounts of data provide little value if they cannot be turned into actionable information. Reliability teams have significantly more data to analyze, but they have fewer analysts to accomplish that work. Analysts have an increasingly niche, specialized and expensive skillset, making them hard to find, and even harder to retain. Consequently, many plants have a massive investment in data collection infrastructure, without an efficient way to turn this data into actionable information. 

Remote analysis is the solution to this problem. Teams implementing a remote analysis solution send their data to analytics experts. If these experts are internal, they can provide analysis for multiple sites. If they are external, they typically provide analysis for a wide variety of organizations.  

A team can choose to do this with all their data, or even use such a solution to augment the analysis they are already performing onsite. As reliability teams evaluate remote analytics solutions, they can significantly improve their chances of success by following four key strategies. 

1: Collect the right data 

Teams that have not implemented extensive wireless monitoring solutions are not excluded from remote analytics solutions. It is possible to collect data for a remote analytics program using scheduled manual rounds with a handheld analyzer. Teams collecting data via manual rounds must ensure they collect consistent data. For example, if a technician is collecting data on “motor outboard horizontal,” the data must be collected from the same spot each time, a task that can be particularly difficult to accomplish if several different technicians alternate on the route.  

One strategy to ensure consistent collection of data is to mark the location for collection on each asset to remove guesswork. However, an even better strategy is to create comprehensive training to ensure every technician knows exactly how to collect data—without significant variation—on every asset in the plant. Such a training plan should be well documented so new technicians can be brought up to speed as quickly as possible. 

For teams using wireless condition monitoring, collecting data is significantly easier. With wireless triaxial sensors, as long as the device is installed at the right location—typically as close to the bearing as possible—teams will know they are getting the right reading each time. In addition, wireless sensors ensure that readings are taken where they are needed without fail and ensure consistency of testing schedule. Regardless of how busy the staff is on a given day or who is out of the plant, the right data will come in at the right time. 

One other key step in collecting the right data is performing a criticality assessment to ensure the team is receiving the correct data. Safety-critical assets and those that interrupt production will be the most important and will likely be covered by more complex systems. But for balance of plant assets, identifying which systems have spares, which have the most impact on production, which are the costliest to repair, and more can help teams determine data collection frequency for each machine to ensure they are not overloaded with data or taxing the wireless network. 

2: Have the right team perform analysis 

Assuming remote analysis cannot be performed in house due to a lack of qualified personnel, the next step is choosing a provider for remote analysis, and it is critical to have experienced people review the plant’s data. The best vibration analysts carry category 3 or 4 certification from the International Organization for Standardization (ISO). ISO category 3 or 4 analysts will have years of experience and will be more likely to have seen a wide variety of issues, better preparing them to identify root cause of the most complex problems.  

In a typical process manufacturing plant, 80% or more of the day-to-day asset problems will likely be the most common ones: balance, misalignment, under-lubrication, etc. These problems will be easy for nearly any analyst to identify and isolate, and then help plant personnel with resolution. However, the other 20% of problems, due to their complexity, will typically be far more time-consuming to analyze, and will require deep expertise for resolution. A reliability team using remote analysis to drive efficiency needs someone who can solve such problems quickly.  

Highly experienced personnel who have worked in multiple industries with multiple global customers will have seen many more unusual problems than other analysts. They will diagnose problems more quickly, and they will be far more qualified to help teams identify the severity of a problem and whether it needs to be fixed today, or can wait a month, or even a year, for a scheduled outage. 

Moreover, highly experienced, certified analysts will be better prepared to perform the complex multivariate analysis necessary to uncover the most complex problems. Analysts primarily use vibration to identify issues, but the best providers will also be able to work with the plant’s reliability team to check other process variables to help discover root cause when necessary. If something in the process changes, it can have a significant impact on asset reliability. An ISO category 3 or 4 analyst will have a much easier time using the available data to identify those changes.

3: Insist on meaningful reports 

The most meaningful reports from a remote analytics group will identify problems and do so in a way that fits the needs of a wide variety of stakeholders. Teams need different reports for each role. For example: 

• One that an onsite analyst can examine and evaluate. 
• One that a technician can look at to guide his or her actions. 
• One that a manager can explore to track and trend reliability and performance. 

First and foremost, analytics reports should quickly and easily identify problems and draw focus to what needs to be done first. The best reports provide a simple “green, yellow, red” view of asset status to quickly show teams of any experience level how to prioritize their action. Assets in green are healthy and require no action. Yellow assets have developing problems that should be addressed when teams have availability. Red assets are experiencing critical failures and should be addressed immediately. 

A high-quality analytics report also offers a single page overview of the plant as a whole, showing teams how many assets they have, how many are in each state, and how many are in other stages. Provided with such a view, any reliability team, regardless of size or experience level, can easily maintain a holistic view of plant health.  

Teams also need access to deeper analytic details, such as spectrum, waveform, impacting or other variables necessary to identify problems. While not every team will want or need such deep data, plants with onsite analysts will require access to the information they need to make key decisions about their assets. 

Perhaps most importantly, reports should not just list asset health, but also provide actionable information to resolve problems, and the reasoning behind those decisions. Teams should be able to read a report and know exactly what to do to solve the problems that must be addressed.  

Moreover, they should also have a way to provide feedback so they can report what they find when they perform repairs. Armed with that information, remote analysts can close the loop on problems, ensuring corrective actions resulted in expected outcomes. 

4: Find a solution that can be customized to the plant’s unique needs 

No high-quality analytics solution is going to be one-size-fits-all. Every plant is unique and uses different equipment, technologies and personnel to operate at its best. As a result, the best remote analytics solutions are the ones that can be adjusted to meet a plant’s specific needs. 

For example, every plant handles cybersecurity in its own way, and every plant will want to protect its data. Finding a remote analytics provider who can work with the plant’s unique defense-in-depth strategy to make their solution work will lead to much more positive outcomes. 

In addition, with the rising use of artificial intelligence (AI) in analytics, teams will want to identify a solution with the right mix of human analysis and AI solutions. Many organizations use AI to do some or most of their analysis. But while the pattern recognition strategies in AI are useful for fast results, they can also frequently misidentify faults. Reliability teams will get better results from providers who also use certified human experts to dive into results from AI analysis of raw or contextualized data.  

For example, fault severity is very difficult for AI to distinguish. An AI solution can identify the same pattern on two different bearings, one that may last another month, and one that might last a year. In most cases, it takes an ISO certified analyst with years of experience to truly know the difference. 

Knowledge informs the best solutions  

As reliability teams struggle to do more with less, many need outside help, either to supplement the analytics work they are already doing, or to close gaps created by personnel shortages. Such solutions can help teams drive higher efficiency across the plant to secure competitive advantage and be easily customized to meet their unique needs. Finding the right remote analytics solution is not difficult, it just requires knowing the benchmarks to look for and asking the right questions.  

_{All figures courtesy of Emerson}