New Trends in Remote Monitoring Enhance Manufacturing
| By: Kristin Lewotsky, Contributing Editor
In manufacturing, the math is simple: the greater the amount of unscheduled downtime, the lower the throughput for the production line. If a company is not making product, it is losing revenue. Remote monitoring provides a powerful tool to prevent these situations (see figure 1). With continuous updates on asset health, organizations can evolve from reactive maintenance to predictive maintenance, addressing problems before they happen. Along the way, they can reduce cost of ownership and increase operational equipment effectiveness.
Remote condition monitoring has been available in increasingly sophisticated forms for a decade. For much of that time, industry watchers have been predicting that the breakout year was upon us. Somehow, that never quite materialized. Today, the situation may finally be poised for change. Altered staffing needs, coupled with economic trends and a range of new technology offerings may finally create the conditions for broad uptake of remote monitoring.
The changing workforce
Today’s factory workforce looks very different than it did a decade ago. For starters, the effects of the economic downturn forced companies into lean staffing models. Most organizations do not have anything close to the same depth of engineering talent on staff as previously, whether that is in their design team or their maintenance group.
It isn’t just a question of numbers. The demographics of the factory floor have changed. The baby boomer workforce is retiring and taking its expertise along with it. The engineers replacing them lack the decades of hands-on experience. At the same time, the new staffs are challenged by a combination of aging legacy equipment and increasingly complex new equipment with an ongoing pressure to deliver more - more throughput, more quality product, and more uptime.
These factors are pressuring manufacturers into considering remote monitoring, whether they want to or not. “Certain companies we've spoken to have said, ‘Well, we don't have a choice, we have to adopt some of these technologies because we just don't have the manpower to meet all of the maintenance requirements that we have at the moment,’” says Alex West, principal analyst at IHS Manufacturing. “I think sometimes necessity breeds innovation and certain companies are in the position where they're going to have to move. They don't have that option not to do so.”
The latest generation of remote condition monitoring equipment provides resources to bridge the skills gap. Users can choose from budget friendly, turnkey systems to cloud-based solutions supported by off-site experts. “It allows users to access that subject matter expertise as a service,” says Phil Bush, project manager for Rockwell Automation Remote Monitoring and Analytic Services. “We can help customers refine and maximize their usage of their equipment.”
The changing economics make the technology feasible for a much broader range of assets. “Five years ago, vibration monitoring for a motor would cost thousands of dollars,” says Vinicius Strey, digitalization consultant within Siemens Industry Motion Control Services Group. “Now you can implement a system for hundreds of dollars.”
Applying remote monitoring
Condition-monitoring systems track parameters like current draw, temperature, and vibration. In some cases, the sensors are embedded in the components themselves. Many of today’s smart drives include current sensors that monitor the motor and temperature sensors that monitor both motor and drive. Encoders can include thermocouples, as well.
If a motor requires more current torque to move the load, it could indicate a bearing problem or a jam. An increase in temperature can provide advance notice of lubricant breakdown. If the resonant frequency of a belt-driven actuator increases, it could be a sign that the belt is over tensioned. If the resonant frequency drops, it could mean that the belt is loose.
Today’s smart drives can track asset health using model-based monitoring. The motor tries to position the load by applying torque. The feedback loop reports error, which triggers a torque adjustment. This change introduces vibration. Based on torque adjustments, the drive uses an embedded model to estimate vibration; it performs a similar exercise for friction. “Vibration happens when the torque is applied dramatically and adjusted massively,” says Elaine Wang, senior product marketing engineer at Mitsubishi Electric Automation (Vernon Hills, Illinois). “in this way, the servo amplifier itself can monitor the changes in torque and then, based on servo reactions, it estimates vibration.”
Some smart components get quite sophisticated, monitoring inrush current, for example. In other cases, based on calculated vibration and friction coupled with total travel distance, the drive can predict the point of failure of the motor.
Other systems monitor equipment based on actual readings. Analysis of this data can generate very granular feedback on system health. The technique is applied to the output from a three-axis accelerometer mounted on a rotating shaft. Data can be presented in a number of ways, including as a time waveform and as a frequency spectrum (see figure 2).
The largest peak will appear at the fundamental frequency of the rotating shaft. Additional peaks appear at harmonics of the fundamental frequency (e.g. 2X, 3X, etc.). These so-called synchronous peaks correlate to vibration related to shaft rotation, such as the vanes of a fan or pump. Peaks that appear below the fundamental frequency are known as sub-synchronous peaks. They correspond to vibration caused by sub-synchronous motion such as from loose belts, bearing cage defects, etc.
Each device has a characteristic vibration signature. Changes to peak amplitudes or the appearance of new harmonics indicate alterations in the physical condition of the asset (see figure 3).
By monitoring the vibration spectrum for change, it is possible to discover and monitor signs of wear or damage quite precisely. Just by comparing a vibration spectrum to baseline, a trained vibration analyst can detect the presence of a bearing defect, a broken impeller blade, and much more (see figure 4).
Although these techniques may be more commonly associated with fixed-speed AC induction motors, they can also be used with induction motors operated with variable-frequency drives and with servomotors. In the case of servo motors, the vibration frequencies simply occur at higher frequencies; as long as there is a baseline for comparison, vibration analysis can still reveal changes and problems.
Running equipment to failure can have disastrous consequences. In the best case scenario, a bearing, for example, fails and needs to be replaced. In the worst case scenario, failure can damage the motor shaft, couplings, or even the gearbox. The repair process changes from simply swapping out a component to ordering a new motor, having parts custom machined, etc. Depending upon the location of the equipment, additional time may be required to remove the damaged parts and put in new ones. Downtime can stretch from minutes to hours, or even days, with costs mounting all the time.
There are four types of assets that can benefit from remote condition monitoring:
- High-value assets - sophisticated equipment that is time-consuming and expensive to repair or replace. Examples include gas turbines and generators. For many years, the cost and complexity of condition monitoring systems, not to mention the effort required for integration, typically restricted their use to only high-value assets.
- Troubled assets - equipment that is operational but demonstrating damage signatures. For instances in which repairing the unit would require a line stoppage, asset owners may want to run the asset through to the next maintenance period, if possible. Online condition monitoring can be set to analyze the performance of the asset 24/7, automatically sending an alert as soon as the metrics exceed a preset threshold.
- Critical assets - equipment whose failure can shut down production. Even a motor costing a few hundred dollars can be a critical asset if its failure prevents the line from operating. Online condition monitoring can be set up to transmit alerts any time the performance of the equipment deviates from normal. Maintenance can examine the issue and make the decision to stop for repairs for continue running.
- Hard-to-reach assets - equipment that is inconvenient, unsafe, or even impossible to access for standard maintenance checks. Online condition monitoring makes it possible to track the health of the components without putting staff at risk.
Online condition monitoring makes it possible to track asset health while scheduling maintenance at an appropriate interval. The data gathered during this monitoring also makes it possible to achieve a good understanding of the performance and evolution of the asset over its lifetime. This information can be used to develop a model for predictive maintenance on how long that troubled asset can safely run. “If you have a decent platform that can run some kind of trend analysis, you can say hey, I will reach my breakpoint in six months so you have enough time to set up a date for maintenance when production will be low,” says Strey.
Remote monitoring enables organizations to establish service-level agreements with their customers. This is part of a larger trend toward the equipment manufacturers maintaining "fleets" of machines that they operate at the asset owner facilities. The machine builder/OEM is not paid for the equipment but for uptime. Remote monitoring is an essential part of this model because it enables the OEM to deliver the product – uptime.
Paths to success
Strey points to two key criteria for determining whether an asset is appropriate for remote monitoring. First, is remote monitoring technically feasible? Does the component incorporate any kind of sensors or model-based systems? It does no good to have an analytical package with a customized dashboard if the equipment cannot provide the basic data required to monitor the system. Assets recently manufactured typically have this functionality while legacy equipment built 30 years ago may not.
The second criterion encompasses the data that provides a historical context for understanding the results. Start with the low hanging fruit, the motors or machines that have caused issues in the past - this equipments' known failure modes and a record of behavior. The combination of the two sets the stage for moving to remote monitoring.
In this instance, success breeds success. “You can justify the investment going forward by monitoring those assets,” Strey says. “With that success, you can expand your program and roll out across the factory. That's the best strategy. We don't recommend that customers start a big bang remote maintenance program and then connect all the factory at the same time. They should always use a phased approach.”
A phased approach is particularly important for facilities involving equipment from multiple vendors. For a factory built around a single technology platform, implementing an enterprise-wide rollout of remote monitoring for a single technology platform is at least theoretically feasible. Many facilities, however, may choose one vendor for motors and drives, another for machine tools, yet another for controls platforms and SCADA systems. The problem is that each vendor may have its own remote monitoring application that will not communicate with others. Manufacturers with ambitions of rolling all of their remote monitoring activities into a single enterprise-wide application face significant challenges.
“You can achieve success by doing that but most of the time you will be able to get only a shell of the information you want to acquire. You will not be able to get to the deep layers of your machine or your assets. All these vendors, they have their own technology stacks, applications to connect to their remote based solutions, and remote monitoring solutions, so it is a tough approach if you use all those various solutions - but you do get the best information out of your assets. Again, it is harder to manage than if you define one solution for your whole company. That's an important aspect. I have seen both sides, both work but you have different challenges to tackle.”
The optimal approach to take varies depending upon the particulars of the equipment involved. If the data is already being collected in some form at the customer site, then capturing it and sending it to the cloud for analysis is a straightforward next step.
Asset owners and OEMs increasingly can access analytics suites to convert data into actionable information. This can take place on a cloud basis or locally on premises. As with most things in engineering, the optimal choice depends upon the details of the application. “Not everything needs to go to the cloud,” says Bush. “In fact, in a lot of ways the cloud can be more expensive than doing it locally. Understanding where in your information space the data is appropriate or the analytics is appropriate is part of part of the evaluation.”
The most likely scenario is a mix of assets, some a few years old and some several decades old. Each will have different levels of instrumentation and will be data intensive to a different degree. In this case, the first step is simply to begin collecting data and processing it locally. Once success has been demonstrated, the next step is to move toward edge processing, then toward the connected enterprise and cloud-based processing.
"It's kind of a crawl, walk, run type of activity because going from no instrumentation to fully instrumented in the cloud, is a significant investment,” says Bush. “It's really about picking the high-value or high-volume items, attacking those, getting some return on those efforts, and then rolling that back in as a future investment, to go do the next hot item on the list."
Getting to yes
Cost is one of the barriers to entry, although the latest generation of economical solutions promises to change that. “When we talk to customers who don’t have exposure to condition monitoring, they always think it is very expensive,” says Abhishek Jadhav, Product Manager, Dynapar (Gurnee, Illinois). “Once they take a look, they realize hey, it's not.”
Responsibility also falls to the vendors to create user-friendly platforms. They can also assist with changing the corporate culture, says West. “Many vendors of these types of solutions are building in consultancy into their offerings. To a lesser degree, some companies are even supporting on the change management. “
Increasing adoption requires participation at multiple levels. Within organizations, management needs to endorse the idea and support the transition to remote monitoring. That’s an increasing trend, as the decision-making process continues to evolve from silo-based decisions within engineering, purchasing, and maintenance to organization-wide initiatives. “Five years ago, we didn't have the upper management support or C level push to implement these solutions,” says Strey. “Basically we had to talk to the maintenance manager and the maintenance manager would try to sell that internally. Now, upper management is going to conferences where they are being taught about these solutions. They are getting back to their maintenance staff and saying, ‘Hey, we've got to digitalize our factories. We've got to implement predictive maintenance.’ I think that is the key change going forward.”
The perennial question is when, exactly, remote monitoring will takeoff. Here, opinions vary somewhat. “We are already seeing a tremendous amount of customer interest,” says Bush. “What is going to limit the actual implementation is really around how many assets are actually instrumented and connected. I think you will see a large break out with those customers that are connected but I think the majority of the industrial automation space is still trying to get to that point where they are ready to jump into this digitally enabled space. They are going to have to take the next year or two to make an investment to position themselves to take advantage of that.”
“I certainly think we're still very much in the trial phase and then the roll out phase at the moment with many companies,” says West. “There are others that are much more traditional and still very worried about opening up their network to a third party. That's really going to hinder the evolution of these products within some organizations.
Jadhav disagrees. “I think it is going to be a huge boom next year, and exponentially change too.”