Nothing grabs the attention of top management like talking about profits, which explains some of the current buzz around predictive maintenance.
For decades, geeks in heavy industry have been exploring the field of predictive maintenance in depth. The basic idea is simple:By combining the right mathematical analysis with sensor data, machines can be repaired before they break down, saving valuable resources and eliminating unplanned downtime. But for a long time, the processes that made predictive maintenance possible either did not exist or were too expensive to be considered practical.
Things look very different today, with the combination of advanced analytics, low-cost sensors and the Internet of Things (IoT) promising to elevate maintenance from a cost center to a profit center. True believers see predictive maintenance as the spark that can ignite an economic revolution, and from their perspective, wider adoption of predictive maintenance principles will enable companies to offer a much wider range of products and services than ever before. Early adopters are likely to be companies in the energy, transportation, manufacturing and information technology sectors. As more and more sectors of the economy become dependent on services and revenues flowing through the IoT, the appeal of predictive maintenance will expand rapidly.
Let's look at the data first.
Authoritative studies have shown that a Functional Predictive Maintenance Program can reduce maintenance costs by 301 TP3T, reduce downtime by 451 TP3T, and eliminate failures by as much as 75 1 TP3T.
The advantages of predictive maintenance are numerous. A well-planned predictive maintenance program can virtually eliminate catastrophic equipment failures. We will be able to schedule maintenance activities to minimize or eliminate overtime costs. We will be able to minimize inventory and order parts in advance as needed to support downstream maintenance needs. We can optimize equipment operations, save energy costs and improve plant reliability.
If the maintenance staff makes the leap from the garage to top management, it will follow the path pioneered years ago by monotonous back-office functions such as accounting, which evolved into finance and were led by the chief financial officer (CFO); similarly, data processing evolved into IT and was led by the chief information officer (CIO).
Is this the dawn of the predictive maintenance era?Mike Hitmar, global product marketing manager at SAS, who specializes in manufacturing, gives a "resounding yes" to that question. "Data analytics is cool right now, and people are starting to better understand what data analytics can do," he says. "Analytics is the other side of BI (business intelligence). Instead of looking back at what has happened, you can look forward and predict what might happen."
The economic potential of predictive maintenance will fuel the steady growth of the IoT, Hitmar said.Companies like GE, Cisco, IBM and Intel are counting on the predictive maintenance capabilities enabled by the IoT to create at least $100 billion in additional value for the energy and utilities industries, Hitmar said. According to Gartner, the IoT is projected to create nearly $2 trillion in value, with new value and much of the value creation in the global economy over the next five years to be fueled by predictive maintenance.
Ganesh Bell, Chief Digital Officer and Managing Director of GE Power & Water, believes that a predictive maintenance strategy can create three levels of value, as shown in the table.
| organizational level | recount (e.g. results of election) | goal |
|---|---|---|
| Top management | Market performance | Optimizing corporate profitability |
| deputy director (of a company etc) | Operations Optimization | Improve efficiency; reduce overall operating costs |
| managerial staff | asset performance | Increased asset reliability and availability; reduced maintenance costs |
Describing the first layer, he said:
For plant managers and maintenance managers, the primary focus is on asset performance. Their goal is to have zero unplanned downtime for every asset they own. That's the foundation - improving asset reliability and availability while reducing maintenance costs.
Bell adds, "The next tier is the vice presidents, who look at optimizing the entire plant operation, not just the physical assets, but everything, including the supply chain and human resources."
The third and highest tier - top management - focuses on optimizing profitability across the enterprise. "In the energy sector, when we talk about saving 1% of fuel, it's going to be worth about $65 billion to our customers," he says. "From our perspective, we've seen predictive maintenance have a significant business impact at all three levels."
Each layer has different perspectives and goals. At the bottom layer, managers and operators must understand the physical characteristics of individual parts and machines. At the next level, the interaction between resources, machines, processes and human behavior is critical. At the top layer, the focus is on ensuring that the efficiencies realized at the lower layers add to market advantage and actual profits.
Clearly, predictive maintenance is more than just a tool or a solution; it is an integrated business strategy with multiple layers, interconnected processes, and complex relationships between a variety of stakeholders across organizations and beyond traditional boundaries.
Bell says he sees parallels between the evolution of predictive maintenance and the evolution of ERP (enterprise resource planning), CRM (customer relationship management), supply chain management and other systems that have become integral and important parts of an organization's IT portfolio. "We've seen CIOs get involved and work with their company's asset owners or operations leaders to build the IT infrastructure necessary to support predictive maintenance," he says.
He added that as CIOs prepare IT departments for the move to predictive maintenance, it's important to move beyond their traditional roles as "digitizers". "Predictive maintenance is not like replacing atoms with electrons or using software to execute business processes. It's something fundamentally different; it's about creating new value and new revenue for the company."
Prevention and forecasting
Preventive maintenancerespond in singingPredictive maintenanceThe difference between is not just semantic. Imagine a three-tiered pyramid. At the bottom is theReactive maintenanceThe operating philosophy is "wait until it breaks, then fix it". The next level is preventive maintenance, where repairs or modifications are carried out at predetermined intervals. The goal of preventive maintenance is to extend the life of the machine and its components.
At the top of our imaginary pyramid is predictive maintenance, which attempts to avoid problems before they actually occur. In a predictive maintenance scenario, the goal is to eliminate unplanned outages or failures altogether. It's not hard to see why utilities are leading the way in predictive maintenance: outages are costly to remediate, pose all sorts of real dangers, and are sure to anger customers. For similar reasons, medical device manufacturers are pioneering predictive maintenance.
Based on information provided by DOE, Tables 2 and 3 show the major differences between preventive and predictive maintenance.
| Preventive maintenance | Predictive maintenance |
|---|---|
| Cost-effective in many capital-intensive processes | Increased component life/availability |
| Flexibility allows adjustment of maintenance intervals | Allowing pre-emptive corrective measures |
| Add component life cycle | Reduce equipment or process downtime |
| energy conservation | Lower parts and labor costs |
| Reduced equipment or process failures | Better product quality |
| Estimated savings over reactive maintenance program 12% to 18% | Improving worker and environmental safety |
| Improvement of workers' morale | |
| energy conservation | |
| Projected savings over preventive maintenance program 8% to 12% |
| Preventive maintenance | Predictive maintenance |
|---|---|
| Catastrophic failures are still possible | Increased investment in diagnostic equipment |
| labor-intensive | Increase investment in staff training |
| Performance including unnecessary maintenance | Savings potential not easily seen by management |
| May cause accidental damage to components during unnecessary maintenance |
Follow the money
Greg Fell is the former CIO of heavy equipment manufacturer Terex, and prior to that he held a technical management position at Ford Motor Co. Fell believes the practical and economic arguments in favor of predictive maintenance have become too strong to ignore.
"The best way to think about predictive maintenance is to tie it to a revenue stream," he says:
When your machine is up and running, you are making money. When your machine breaks down, you are losing money. A typical automobile manufacturer produces one car every 60 seconds. If each car retails for $40,000, your gains or losses can add up very quickly.
While preventive maintenance relies on the simple concept of "Mean Time Between Failures" to create a practical maintenance program, predictive maintenance is based on a deeper and more fundamental understanding of the physics behind the operation of a machine and its individual components.
"Instead of just looking at mean time between failures, you're looking for subtle clues within the machine itself," Fell says. "You're measuring sound, heat, vibration, tilt, acceleration, compression, humidity, and checking to see if any of them are out of spec." Fell adds:
The basic idea of predictive maintenance is not new. What has changed is that it is much less expensive to get data from machines today than it was in the past. Twenty years ago, an accelerometer cost thousands of dollars. Today, every smartphone has one built in. The technology needed for predictive maintenance has been miniaturized and the cost has dropped dramatically.
The cost of transferring data from machines to data repositories has also decreased. In the past, machine-generated operational data was collected manually by technicians on the shop floor. Today, this data can be sent wirelessly to the Internet via Bluetooth or Wi-Fi.
Not all jobs are created equal
Another problem facing traditional preventive maintenance programs is the assumption that every machine of a particular type will operate under similar conditions or with similar parameters.Clifton Triplett is a managing partner at SteelePointe Partners, a management consulting firm. He is also the former Chief Information Officer of Baker Hughes, a $20 billion global oilfield services company, and head of manufacturing processes at General Motors. As a West Point graduate, Triplett understands that equipment often operates in situations and environments not anticipated by design engineers or manufacturers.
"It's important to remember that not all jobs are created equal," says Triplett. "If you're running a tool within the operating limits for which it was designed, it will require some level of maintenance. But if you're running it outdoors, well below or above normal operating specifications, this is a time when a different level of service will usually be optimal."
For example, a truck primarily used to transport ore in Canada will require a different level of service if it is used to haul ore back and forth up a mountain, on a flat asphalt road, or in a hot, dusty desert area. In the oil and gas industry, drill bits primarily used in conventional "straight down" drilling operations require different reliability parameters and service requirements than those used in more challenging unconventional horizontal drilling.Reliability is partly defined by design, but reusable equipment depends heavily on the maintenance services performed on it.
"In the Army, we run tools and equipment in as many different scenarios as possible to see how they perform and react in different environments," Triplett said:
We will test the equipment at different altitudes, temperatures and humidity. We will test equipment to see if fresh or salt water affects its performance. Whatever comes to mind, we have tried and evaluated how to adapt our maintenance programs to the conditions our equipment is exposed to.
Companies that emulate the military's focus on maintenance issues can charge extra for their services. For example, as drilling conditions become more extreme, the value of reliability increases accordingly. If you're an independent oil driller and you have a good predictive maintenance program (assuming it's been proven by reliability metrics and performance), you may demand higher service fees or accept more challenging opportunities, Triplett says. non-productive time (NPT) is "an evil that all drilling companies seek to eliminate". ". Service or operational failures are the number one contributor to NPT, and trust in a service company's maintenance program builds customer confidence when awarding work.
"If the oilfield operator trusts your maintenance program, they're also more likely to let you use your existing equipment longer because they won't be as concerned about it failing," Triplett says. "If you have a poor maintenance strategy or track record, they are more likely to require you to use new equipment every time you start drilling. Being forced to always use new tools is very costly and likely not competitive."
Smart operators understand that predictive maintenance translates into pricing power, and now, as drilling technology becomes more sophisticated, the opportunity for market share is growing. Halliburton, for example, has a reputation for being able to operate reliably in hot environments longer than its competitors. This reputation creates pricing power and generates higher profits.
lay the foundation
Daniel Koffler is chief technology officer at Rio Tinto Alcan (RTA), a global leader in the aluminum business and one of five product groups operated by the Rio Tinto Group, a multinational metals and mining company, which produces 31.4 million tonnes of bauxite, 7 million tonnes of alumina (aluminium oxide), and 2.2 million tonnes of primary aluminium annually. Koffler is responsible for keeping RTA's machinery running.
"There is inherent downtime in any kind of failure operating scenario. When your primary asset fails, you're forced to keep additional assets to make up for the loss," Koffler said. "Either way, you lose production capacity during repairs and spend money on idle additional assets."
While scheduled maintenance models can prevent unplanned downtime, there's also a good chance you'll end up repairing equipment or replacing parts unnecessarily. "With predictive maintenance modeling, we can keep assets running longermergeAvoid unplanned downtime," Koffler said.
According to Koffler, reliable data and solid computational models are the foundation of predictive maintenance. In addition, the corporate culture must adapt to processes that don't always produce perfect results.
"At management level, people need to accept that data modeling doesn't start with 100% maturity. It's a process that takes time. You don't start at peak," Koffler says. "That means you have to accept additional risk. There may be unexpected failures that are part of the process."
Koffler emphasizes that predictive maintenance is not a magic formula; it's an iterative, scientific process that needs to mature over time:
You need collaboration between subject matter experts such as mechanics and data scientists. Mechanics understand how machines work and data scientists know how to build data models. The mechanic's knowledge should be encoded into the data model. Mechanics and data scientists need to communicate over time to refine the model. This cannot be done without cross-functional collaboration.
As with any process based on statistical analysis, predictive maintenance is inherently imprecise. The risks associated with predictive strategies must be "negotiated" and understood by the relevant stakeholders, he said:
You may want to take an extremely conservative approach and replace the part earlier, but still maximize value by pushing the part to its predicted point of failure. It will then become a discussion about risk rather than analysis. These discussions about cost, productivity and risk happen every day in business. It is the nature of real life.
Koffler doesn't believe predictive analytics will become a mainstream consumer product in the near future. "You have to do a cost-benefit analysis. Putting in sensors, collecting data, analyzing data - it all costs money," he says. "Just because you can do something doesn't mean you have to. Take the belt in a car engine, for example. You can put sensors on the belt, but it's more economical to run them until they break and then replace them."
Not just heavy machinery
Not every aspect of predictive maintenance revolves around heavy machinery and industrial processes.Doug Sauder leads the research and development team at Precision Planting, which develops technology to help farmers improve seed spacing, depth and rooting in their fields.
"The agricultural challenges of the future are all about doing more with less in a sustainable way," Sauder said. "It's about meeting the needs of a growing population and being environmentally responsible." In his view, predictive analytics play an absolutely critical role in any sensible solution.
"We can be smarter about maximizing every square inch of land, from what seeds to plant, how many seeds, seeding correctly, watering and applying the right amount of fertilizer," Sauder said:
For example, it is important to model nitrogen in fields. You need to model rainfall and understand how it disperses into the ground. You need to have the ability to predict where nitrogen will move so you know when to apply more nitrogen to the field.
Sauder and his team not only help farmers learn more about their fields; they also "train" farm equipment to perform better. "We call it 'smart ironing,' and it's essentially a technology that allows a farmer's equipment to think independently as it travels across the field," he says.
From the air, all cornfields look very similar. But on the ground, each field is unique. As a corn planter moves through a field, it constantly encounters a variety of soil conditions. Some soils are hard; others are soft. If the seeding pressure on the equipment is static, some seeds will be deposited deeper than others, resulting in uneven growth. But farmers won't realize the problem until months later, and by then it's too late to fix it.
Sauder said:
We place sensors on the planter and take hundreds of measurements per second. We can change the pressure in real time as the planter moves across the field to make sure every seed is in the optimal position. With the 'smart iron' we can literally micromanage every bit of the field and make sure that every seed is placed in the right environment.
The ability to "micromanage" a cornfield requires the same combination of predictive capabilities needed to manage the performance of gas turbines, jet engines, nuclear generators, internal combustion locomotives, ore haulers, and magnetic resonance imaging equipment. As Koffler mentioned earlier in this report, predictive maintenance is a multidisciplinary science with roots and branches far beyond heavy industry.
The future of maintenance
The next step in the evolution of predictive maintenance may involve a greater amount of machine learning and closed-loop automation techniques. Currently, predictive maintenance systems are limited to signaling hazards and issuing alerts about impeding failures. Future versions will undoubtedly include advanced decision-making tools and recommendation engines.
Prakash Seshadri, head of product development at Mu Sigma, one of the world's largest decision science and analytics companies, sees predictive maintenance inevitably evolving into prescriptive maintenance. In a prescriptive maintenance scenario, the system won't just tell you that something bad is about to happen; it will provide useful advice.
Seshadri says, "It's not just about saying, 'This has the potential to break down,'"
A predictive maintenance system will say, "This is what you should do based on current and expected conditions" and offer a range of choices that will guide the person to make better decisions. But at the same time, if the human decides to override the advice, the system will capture the behavior and evolve through learning.
Today, predictive maintenance is widely used in the information technology, manufacturing, healthcare and energy sectors. In the near future, predictive maintenance will be more widely used in the retail, telecommunications, media, and financial industries." The potential for "cross-infection" seems limitless.
English 
Chinese 
German
Sounds interesting.