Condition-based monitoring is the most effective approach to machine maintenance. A reactive maintenance approach only comes in after something has already gone wrong. However, a preventive approach happens at pre-scheduled intervals with condition-based monitoring giving maintenance managers oversight of machines at all times. This way, they can send in technicians only when it’s necessary and before an issue leads to machine downtime. 

A version of this post was originally published on May 4th, 2021.

Fitbit your Factory

Condition-based monitoring is like putting a health tracker on a piece of industrial equipment. Sensors attached to equipment assess aspects of machine health through various methods – ultrasound, thermography, specialized electrical testing, oil analysis, and vibration analysis. In turn,  maintenance teams can use that data to identify patterns and predict machine failure. As a result, three essential workflow transformation occur:

• Reactive to proactive maintenance
• Negative to positive reward reinforcement 
• Open- to closed-looped controls

Vibration analysis is an especially important form of condition-based monitoring because it provides the deepest insights into a machine. By monitoring vibration through sensors, manufacturers can note changes to detect the full spectrum of machine faults.

In an earlier era, vibration analysis required expensive equipment, complicated software, skilled analysts, and lots of manual effort to gather and interpret data. That’s why factories knew relatively little about machine health and had to rely on reactive or predictive approaches to maintenance.

Vibration Analysis For All

With the advent of new and innovative technologies, however, vibration analysis and other forms of condition-based monitoring have never been easier, cheaper, or more perceptive than they are today.

What’s more, Augury’s accessible and holistic machine health solutions take maintenance a step beyond predicting failure to providing prescriptive insights. Augury’s sensors perform condition-based monitoring and then feed data into advanced algorithms that accurately diagnose faults and suggest corrective actions. Manufacturers can receive prescriptive insights on every machine.

This accessibility means condition-based monitoring and prescriptive maintenance are no longer the next big things in manufacturing — they’re already here, and soon they will be table stakes.

A Paradigm Shift in Progress

Now that machine health data has become easier than ever before to monitor, analyze, and act upon, maintenance will never be the same. As innovative approaches become the norm in manufacturing, we’re seeing the following paradigm shifts in the industry:

1. Reactive to proactive maintenance

Once manufacturers gain the ability to track machine health in depth and in real time, they don’t have to wait for breakdowns to realize there’s a problem. They know well in advance and can send technicians on-site at the earliest warning sign.

A condition-based monitoring program built on quality machine health data makes proactive maintenance a reality for the first time ever. Manufacturers can act early and effectively so that machine downtime and a host of other issues have less risk of disrupting production.

2. Negative to positive reward reinforcement

A reactive approach to maintenance leads to negative reinforcements. When maintenance teams are always reacting to emergencies, somebody ends up in the hot seat — someone is too slow, unprepared, making mistakes. In this kind of high-stress environment, motivation comes from all the wrong places; people work against each other rather than with each other.

On the contrary, a condition-based monitoring approach to maintenance can encourage the exact opposite kind of reinforcement. Instead of focusing on placing blame for the catastrophe that did happen, you focus on celebrating the successes — or the catastrophes you avoided. For example, if a technician goes in to repair a machine and sees damage that surely would have led to failure, they know they have actually prevented that failure.

It’s easy to envision how expensive and time-consuming the repair would have been. Finding the wins to celebrate is easy with a condition-based approach. Now, instead of being motivated by a fear of failing or being called out, maintenance teams are motivated by the opportunity to achieve their best.

3. Open- to closed-looped controls

Two common control schemes that everyone experiences in their daily lives are open loop and closed loop. Open-loop control means performing a series of predefined actions and hoping they have the desired outcome, though the final outcome is fraught with uncertainty.

If only one ceiling fan needs to be cleaned, for instance, why clean them all?

This is analogous to managing a plan using only planned maintenance. You do maintenance tasks based on a schedule but not based on the unique needs of your equipment. This could mean you miss important maintenance tasks on some items or do unnecessary ones on others. If only one ceiling fan needs to be cleaned, for instance, why clean them all?

Condition-based monitoring and machine health, on the other hand, serve as the feedback mechanisms necessary to enable closed-loop control. This control scheme sets specific targets and measures progress against those targets so fine corrections can be made to keep them on track.

If your target is a healthy machine, for example, you can make repairs until the desired outcome is achieved. If a machine starts to deteriorate, you can identify the problem early and make small corrections to make it healthy again. All precision maintenance programs should be led by closed-loop control.

Condition-Based Monitoring: A Foundation for Transformation

Each of these shifts is pivotal on its own, but together, they form a strong foundation for the transformation of maintenance and reliability best practices. With a condition-based approach and predictive insights, you can leave outdated schedules and unexpected fire drills in the past and lean into the future of maintenance with more data and better insights.

Learn more about enabling condition-based maintenance.

The post Condition-Based Monitoring: The Future of Machine Maintenance appeared first on Augury.

Machines have their own language. When operating, they’re constantly broadcasting signals of their status. To a large extent, machine health is about translating this language to understand where, when, why, and how a machine requires attention or intervention.

That’s where vibration analysis comes in. The source of a vibration, the direction (vertical, horizontal, or axial), and the vibration intensity can all indicate impending problems. For instance, if a shaft is supported by four bearings and one vibrates differently to the others, that bearing may be headed toward failure. If vibration analysts spot the problem in time, they can intervene early. If they don’t, however, preventable problems can turn into catastrophic failures.

Happily, today AI can do more of the heavy lifting while empowering vibration analysts to focus on higher-value work:

• Become more action-oriented
• Transform the culture
• Monitor far and wide

Taking Control of the Vibration Analysis Data

More vibration data translates into better insights about machine health. Unfortunately, the sheer volume and velocity of the data available creates a serious obstacle in vibration analysis. When analysts have vibration data coming from many sources on a constant basis, it becomes an overwhelming effort to understand what it all means.

In this circumstance, the amount of data manufacturers can analyze depends not on the actual amount available but rather on how much data analysts can handle. Much of the data goes uncollected, or unanalyzed, which means important insights go unnoticed.

How can analysts cover more data to gain better insights into machine health? Artificial intelligence in manufacturing is one possible solution.

To learn more about how Augury’s machine health
solutions can boost your business, get in touch today. 

AI Optimizes the Work of Vibration Analysts

AI and machine learning technologies excel at carrying out routine, repetitive tasks at high speeds on a massive scale. For the purposes of vibration analysis, the tools could instantly examine each piece of vibration data and identify whether it’s irrelevant (as much of it is) or an anomaly that requires analysts’ attention.

Such technologies can also handle the diversity of machinery present within manufacturing facilities and across facilities. Instead of being programmed to know what anomalous vibrations look like on every piece of machinery, AI can learn that information by studying the standard vibrations.

Once it understands what the machine should be doing, it can easily identify when a vibration deviates from the norm. This makes AI widely applicable in diverse industrial settings and a perfect fit for vibration analysis.

3 High-Value Tasks Vibration Analysts Can Focus on With AI

AI isn’t a replacement for human vibration analysts, and we shouldn’t expect it to become one. Vibration analysis is nuanced work, and missing a problem can have serious deep consequences. Therefore, human analysts should always have the final judgment call about when or why to shut down a machine.

Armed with AI, however, analysts can make those calls at exactly the right time and with a higher level of accuracy and precision. When they can spend less time collecting and analyzing data, they can also take on more high-value responsibilities in the following ways:

1) Become Action-Oriented

With AI spotting warning signs in the vibration data, analysts can focus their time and attention on developing action plans to respond to those warning signs. They can determine who needs to be involved, how extensive the fix needs to be, the optimal time to schedule repairs, and other important factors. Then, they can present fully formed plans to decision makers who will be more inclined to take swift, decisive action thanks to the depth of the plan. Much gets lost in the gap between analysis and repair. With AI, analysts can bridge that gap and have a meaningful impact on machine health.

2) Transform The Culture

AI-driven vibration analysis represents a fundamental shift away from reactive maintenance and toward predictive and prescriptive maintenance. Responding to issues before they have consequences is an obvious advantage, but it takes some adjustment. Every aspect of maintenance and management needs to adapt by embracing AI and learning how to use it to the fullest. Vibration analysts will need to lead this change and transform the culture around maintenance, which they can do with all the time that AI saves them.

3) Monitor Far and Wide

With AI in the mix, there’s no excuse not to collect vibration data from as many machines as possible. Cost, workload, expertise, and training aren’t limiting factors anymore. AI-driven vibration analysis breaks down huge barriers to effective machine health maintenance because everything gets monitored, not just the machines deemed most important.

Imagine if there was advance warning of every developing machine health issue, a detailed action plan, and a facility-wide commitment to addressing it as quickly and completely as possible. Downtime would become extremely rare. AI-driven vibration analysis makes that vision possible.

The post How AI-Driven Vibration Analysis Enables Higher-Value Maintenance Work appeared first on Augury.

Machine downtime is a widespread problem in the industrial world — 82% of organizations have reported experiencing at least one unplanned outage each year and 45% of which were unable to provide products or services to customers as a result. Today, Machine Health solutions can be applied to make this problem go away. 

This article was originally published on Jun 15, 2020.

If machine downtime is such a drawback on revenue and productivity, why is it still so prevalent? The issue lies in traditional means of machine maintenance. Facilities have relied primarily on two methods in the past.

Reactive Maintenance: Too Little, Too Late

The first is reactive maintenance. When workers notice symptoms of machine problems (such as poor quality outputs or machines breaking altogether), they reactively investigate the issue.

This reactive model is problematic because once workers know that a problem exists, the machines have already failed and product/output has already been affected. Downtime to repair failed or underperforming machines means stalled or halted production lines, reduced revenue, and an overall hit to productivity.

Preventative Maintenance: Too Much, Too Early

The second traditional maintenance method takes a more proactive approach. However it still fails to significantly reduce machine downtime. Preventive maintenance aims to catch problems sooner as workers perform maintenance tasks on a time-based schedule. But most machine failures are difficult to predict using solely operating hours or life cycles as a determining factor of when they need maintenance.

This preventative approach can also mean longer periods of downtime. If a machine begins underperforming shortly after its scheduled maintenance, it’s likely to continue underperforming — and perhaps even fail — before its next scheduled check.

Machine Health as the Key to Reducing Machine Downtime

What’s more, both of the above maintenance methods have been traditionally focused on responding to symptoms of machine underperformance or failure. Instead the focus should be on detecting potential issues before they arise.

In recent years, however, the industrial world has seen an increase in a third type of maintenance focused more on catching problems ahead of time: predictive maintenance. Predictive models use many different methods (including vibration, oil analysis, thermography, and ultrasound data) to predict and prevent failures by early indicators of machine malfunctions.

Predictive Maintenance: Right Amount, Right Time

Predictive maintenance gets us one step closer to solving the downtime problem. However, most facilities are still missing one link that can make the biggest difference: an approach emphasizing machine health. Route based predictive maintenance tools can alert workers of a problem at an early stage. And without continuous monitoring, workers must still come around to check the machines periodically to notice that something has gone amiss.

Even if you have sensors mounted on a machine, if they aren’t advanced enough, they may not automatically diagnose the issue. Hence, these sensors won’t provide maintenance teams the root of the issue, or suggest ways to fix it. So in the end, they’re more like a check engine light.

Confronting The Three Core Causes of Downtime

Machine health, on the other hand, is a framework that elevates predictive maintenance with continuous machine health monitoring. Powered by the IoT and artificial intelligence, this framework can alert workers at the exact moment when issues are detected. It can also provide facility teams prescriptive diagnostics of what’s been detected, and how to fix it.

To get a sense of how machine health could fit into your operation, consider how it addresses the following core causes of machine downtime to keep factories firing on all cylinders:

• Mechanical problems: Sensors perform continuous machine health monitoring to track temperature, vibration, and magnetic data. AI-based algorithms analyze the data and automatically diagnose malfunctions based on changes in this data. Workers are immediately notified of any changes in a machine’s health status with alerts telling them not only why the machine is malfunctioning, but also how to correct the root cause behind the problem.
• Design problems: Continuous machine health monitoring can reveal when individual machines undergo extra amounts of stress due to inefficiencies within the production facility’s design. Smoothing out those points of friction starts by identifying where they are and why they exist, which then leads to facilities that experience fewer issues of all kinds. As a result, facilities can avoid unplanned downtime, delays, bottlenecks and inconsistent product quality.
• Operational problems: Your machine design might be perfect, but humans never will be. As people make modifications to machines, they can unknowingly create additional issues. Machine health technology can create an archive of changes and modifications so you can verify or validate which activities lead to disruptions.

Seize the Day

Unplanned downtime in manufacturing is an inevitable phenomenon. But even planned downtime can hinder your productivity. Nothing can stop these issues altogether, but taking steps to reduce downtime now can have a big impact on your revenue and productivity in the future.

So stop solving problems after they’ve already happened, and start taking a real-time approach to keeping your machines healthy.

Read more about Industry 4.0 Use Cases Enabled by Machine Health.

The post Eliminating Downtime with Machine Health appeared first on Augury.

The blow molder conveyor motor at the bottling plant had been running steadily for years. It drove a belt that moved test-tube-like preforms from the husky to the next stage of the process, where they’d be expanded into plastic bottles.

It wasn’t the most expensive asset on the line, but data confirmed blow moulder was one of the most reliable: its steady pattern of vibrations showed consistent long-term performance. Recently, however, the company’s machine monitoring system tracked a sharp increase in the motor’s vibrations. The blow molder seemed to be suddenly, and rapidly, destabilizing. 

The machine — and the line — were still running. But as the vibrations continued to increase, it was clear they needed to halt production and investigate. Anxiety rose as the line slowed to a stop… they had to locate the malfunction and correct it immediately… or turn the line back on and possibly risk catastrophic failure. If only they knew exactly what was wrong in the first place…

Understanding and Utilizing the Power of Continuous Data Streams 

Advanced digital machine health solutions can interpret continuous data streams and provide actionable maintenance recommendations to protect assets and overall production. Continuous data streams can present great opportunities to increase productivity by—but they can also bog you down if you aren’t able to interpret or utilize the data effectively. 

The secret to dealing with continuous data streams is knowing when to act. The right insights at the right time can mean the difference between missing and exceeding your production goals. If you halt production to inspect an asset and cannot locate the malfunction, you’re running on a double loss of production and labor costs. But with the right amount of advance notice and actionable machine health diagnostics for individual assets, spare parts can be ordered and repairs scheduled during planned downtime, minimizing the impact on production and output. 

Catching Intermittent Conditions 

To determine the ideal time for intervention, reliability professionals can examine a machine’s performance history using data collected on regular intervals. This continuous data stream makes it possible to catch intermittent defects and faults before they become catastrophes. Imagine a sudden fracture of a component, or a chip in a gear, that causes a sudden cliff change in an assets’ vibration data—the response time necessary for action is immediate.

On the other hand, defects caused by wear typically develop slowly and show a slow increase in vibration over time. Some defects take years to materialize, while others can develop over hours, days, or weeks. The rate of change can evolve over time, too, because a defect can develop faster and faster, like a wobbly wheel on a car. The data on these defects show as trends over time. 

In either case, continuous data streams and digital machine health solutions are the best way to prevent and know exactly when to take action and more importantly, what to do to correct the detected malfunction.

Appropriate Response Times

Nobody wants to shut down a line unexpectedly. But In order to maximize the value of continuous monitoring and continuous data streams, response time is critical—especially when corrective actions are necessary. 

Gradual, long-term detection is easier, even for slower algorithms. But fast-developing defects require a real-time data stream because they require a quick response. Without a sophisticated, and fast algorithm, you have less time to react when a machine fails. Even with only five minutes of advance notice, there may be an opportunity to shut the machine down and save the asset or production in some capacity.

Acting too early can also be an issue. Detecting a problem years in advance is possible, but it runs the risk of triggering unnecessary repairs, additional costs, defects in replacement components, or errors in the repair process itself. But don’t forget that any notification that arrives too late is just as bad as no notification at all. The ideal timeframe for when an intervention should occur depends on machine criticality, the detected malfunction, and the industry in which it operates.

Revealing Assets’ Full Life Cycle

Reviewing data collected across an assets’ entire lifecycle allows experts to base future maintenance and repair decisions on the machine’s current and past malfunctions. This helps them plan accordingly and if a machine needs unique components, they can build in time to source the part. Or if a machine, like the bottling plant’s blow molder conveyor motor, has high production value but a low capital investment, they can plan corrective actions as a part of routine maintenance during the next shutdown.

Taking Action

Back at the bottling plant, the technicians had enough time to investigate the truth behind the vibrations: the conveyor motor belt was worn and some of its teeth were breaking off. They emergently ordered a replacement belt and scheduled time to replace the belt shortly after it arrived. Had the plant ignored the alarm, and the belt run to failure, the entire motor could have been compromised, along with the bottling line’s productivity.

Instead, continuous data streams—and a reliable digital machine health solution—demonstrated how actionable machine insights can keep production assets running at their very best, with the right information, at the right time.

Learn more about Machine Health.

The post The Power of Continuous Data Streams: Knowing When to Take Action appeared first on Augury.

There’s no denying that advancements in artificial intelligence and machine learning are occurring faster than most experts anticipated. The era of self driving cars is just on the horizon, the Jabberwacky chatbot can convince you it’s a human, and Google’s deep dream can produce creative works of art. How will AI play out in with predictive maintenance?

Advancements in networking, information processing algorithms, and data storage technologies are enabling computers to acquire complex skillsets and capabilities. In the meantime, the world is left to wonder exactly how these technologies will be implemented and how they will impact existing markets and industries.

The predictive maintenance industry is no exception. There is no question that predictive maintenance is a superior strategy in comparison to common preventive maintenance and especially reactive maintenance. According to the Department of Energy’s operations and maintenance best practices guide, a predictive maintenance strategy can realize savings of 30-40% and 8-12% over reactive and preventive strategies respectively.

The benefits of a predictive program originate from the fact that it is a strategy driven by data such as vibration, ultrasound, temperature, electrical, and other measurements. This data ultimately drives informed decisions such as figuring out the specific repair required to fix a machine or timing the repair appropriately to minimize the risk of catastrophic failure. As such, some of the barriers to the success of a predictive program revolve around the challenges associated with acquiring and managing the vast amounts of information as well as processing or analyzing the information into accurate and actionable results. Not only that, but the industry will face a serious skills shortage if it is to solely use the traditional human-based analysis to keep up with the rising demand for predictive maintenance solutions. Time constraints associated with training proficient human analysts will hinder market growth.

Analyst Trains the Machine

One popular type of machine learning implementation used in machine diagnostics is called supervised learning. In supervised learning, the machine is trained using input datasets that are labeled. For example, if one wants to train a computer to recognize images of cats, they would feed many cat images through the algorithm while telling it that they represent cats. The label specifies the desired output for a given input. This training process can be thought of as building a strong association between pictures of cats and the label “cat.” Then, when a new and unlabeled image of a cat is shown, the algorithm will be able to successfully assign the correct label, and hence correctly classify the image.

With regards to vibration analysis, skilled analysts can play a valuable role in teaching computers how to analyze data. Faults such as shaft misalignment, resonance, imbalance, looseness, bearing wear, and many more, have unique data signatures that indicate the presence of said defects. By analyzing thousands of datasets and labeling them with the faults they represent, professional analysts can provide the input data necessary to teach computers how to identify these faults on their own.

Deep Insights for Human Analysts with Big Data Analysis and Statistical Learning

Not only is there great potential for computer algorithms to learn from human analysts, but humans can learn a lot from computers as well. One of the largest challenges faced by analysts is that mechanical equipment is very diverse. Not only are there many countless types of machines, including chillers, compressors, turbines, fans, and pumps, to name a few, but within each type there are various sub-types and configurations as well. Equipment can be driven by electric motors, gas engines, turbines, and more. There exist various power transmission techniques including direct-driven, belt-driven, and gear-driven. Then the equipment itself can have several unique design details, including the size, materials, geometry, orientation, age, and installation-specific details. All of these variables affect the way a machine vibrates under both normal and faulty conditions.

The prevailing strategy used by many traditional analysts is to blanketly apply simplified tables of limits, such as the ISO 10816 standard, to wide classes of machines. By compiling a massive database of machines that span the full spectrum of design and configuration possibilities, it becomes possible to use big data analytics to gain insights into the expected vibrational characteristics at a more granular level. These insights can augment the analyst’s capabilities, enabling analysts to adjust their techniques on a machine-by-machine basis in order to provide the highest quality analysis possible.

Skills Shortage

A successful predictive maintenance program, especially one that utilizes vibration analysis, requires skilled analysts who know how to interpret vibration data. Becoming proficient as an analyst requires both training and experience. There are four certification levels for vibration analysts which are known as Category I through Category IV. The knowledge, capability, and experience requirements increase as one moves up the ranks. In fact, experience is a required prerequisite for certification eligibility. Due to the fact that experience requires time, it is also a great barrier to market growth. Consider these experience prerequisites of the Vibration Institute, a leading ANSI accredited certifying body:

Minimum Required Work Experience in the Field of Machinery Condition Monitoring and Diagnostics
* Category I: 6 Months
* Category II: 18 Months
* Category III: 36 Months
* Category IV: 60 Months

The highest level of certification, Category IV, requires a whopping 60 months of analysis experience as a bare minimum.

The latest market research estimates that the predictive maintenance market will experience compound annualized growth of 28.4% over the next five years. It is easy to see that the time and training constraints alone prevent human analysts from being capable of keeping up with this demand. A new analyst who begins training today will not be proficient in advanced analysis for at least 3 to 5 years. Meanwhile, the market will have more than tripled in size. Advancements in automated diagnostics are an absolute requirement for enabling the predictive maintenance industry to keep up.

Conclusion

Not only will fusing the diagnostic capabilities of machines and humans enable unprecedented levels of both accuracy and autonomy, it will be absolutely required in order to overcome timing bottlenecks associated with training and certification. Increased automation in diagnostics will act as a backwind for vibration analysts, enabling them to focus more time and attention on the machines with the most severe problems.

The insights gained by taking a big data approach to analysis will improve result quality and analyst skill level. These advancements are breaking down the cost and skill barriers that hinder the implementation of successful predictive maintenance programs. Facilities will have greater access to predictive maintenance, and they will therefore be able to achieve the highest levels of operational excellence and equipment reliability.

Ready for Industry 4.0? Augury PdM can take you there. See how!

The post Diagnostics Fusion: Humans, Machines and the Future of Predictive Maintenance appeared first on Augury.