Which machines can I monitor with predictive maintenance?

This is the introduction to a new series on the types of equipment monitored by AiSight's predictive maintenance solution.

AiSight's predictive maintenance solution provides machine diagnostics and early warning of failures for rotating equipment. That means we monitor almost every machine in every industry, including the textile, automotive, FMCG, wood processing, and metal industries. Our solution uses multiple, powerful sensors, and cutting-edge AI to detect anomalous machine behavior, identify specific problems, and provide warning of breakdowns months in advance. How our predictive maintenance solution identifies machine faults depends on the machine on which it's deployed.

To take a deeper look at how we prevent unplanned downtime, we need to take a look at the functions and uses of each machine—and how our solution monitors each machine and analyzes the data from its sensors. That's why we've launched this series. Over the course of this series, we'll see how AiSight's predictive maintenance detects failures and prevents downtime on everything from LPG ovens to jet dyeing machines.

The predictive maintenance advantage

According to a study in the International Journal of Production Economics, maintenance costs account for 15–40% of production costs (Löfsten, 2000). This is significant, but preferable to the alternative: no production.

Maintenance is important. The better our maintenance practices, the more value we get from our equipment. Keeping machines running yields more production. Keeping machines running longer yields a greater return on investment. And a well-planned maintenance strategy will deliver those benefits at a lower cost.

This is where predictive maintenance comes in. We've written before about how important maintenance planning is—and what an asset a predictive maintenance solution is for maintenance managers. In short, predictive maintenance provides the information maintenance teams need to make airtight maintenance plans, thus reducing overhead, reducing unplanned downtime, reducing unnecessary monitoring and preventative maintenance, and increasing productivity.

This is why we developed our predictive maintenance solution to monitor all rotating equipment.

Which kinds of rotating equipment can I monitor with predictive maintenance?

Monitoring rotating equipment is the best way to provide advanced warning of machine faults. Rotating equipment is nearly ubiquitous in today's production facilities, appearing in critical machines like pumps, extractors, compressors, ventilators, and mixers. Most machines are powered by electric motors—themselves rotating equipment. These may transmit power directly, or through belt drives or gearboxes—more rotating equipment. From there, we often find even more rotating equipment, such as fans or impellers.

Moving parts will all, eventually, break. AiSight's predictive maintenance solution monitors the indicators of breakdowns. Early intervention is best—our analysis of faults finds that 71% of machine fault alerts are caused by relatively minor issues such as dirt accumulation, looseness, and inadequate lubrication. These are simple issues to fix, but cause serious damage when not detected early. And these are only the beginning of what we monitor; our predictive maintenance solution identifies imbalances, misalignments, looseness, bearing defects, friction, gearbox problems, belt drive problems, cavitation, and impacts. How our solution detects and identifies problems depends on the machine.

Before getting into specific machines, it's worth understanding the two fundamental types of rotating equipment: stationary and non-stationary.

What is the difference between stationary and non-stationary rotating equipment?

Stationary rotating equipment runs with fixed operating parameters, including speed, load, and temperature. These machines are either on or off. For example, when we turn on a fan, it runs at one speed until we turn it off. Other examples of stationary rotating equipment include pumps, compressors, and gearboxes—as long as they run at a fixed rotational speed.

Non-stationary rotating equipment runs with variable operating parameters, including speed, load, and temperature. They may, for example, use frequency converters to run at different speeds, as in the case of a bottling machine increasing or decreasing its operating speed. Other examples of non-stationary rotating equipment include turbines and saws, as well as more complicated equipment in servo motors and robotics.

Using predictive maintenance with stationary rotating equipment.

AiSight’s predictive maintenance solution uses vibration analysis to predict faults in stationary rotating equipment. Given that stationary rotating equipment runs at one speed, a properly functioning machine should produce consistent vibrations. The vibration sensors in our Aion sensor nodes detect the vibrations in stationary rotating equipment; our algorithms compare them to normal vibrations, and identify vibrations indicative of faults. Not only that, our root-cause analysis compares the vibrations to other anomalous readings, and identifies the specific problem with the machine.

Vibration analysis makes use of a curve, plotting machine condition relative to time until breakdown. Manual preventative maintenance practices have always used this curve—smoke billowing from a machine is a sign of impending failure, while noise is an early indicator of a problem. Maintenance technicians can get even earlier warning of machine faults through oil particle analysis. AiSight's predictive maintenance solution, using vibration analysis, provides the earliest warning of all—months before a breakdown. This allows maintenance and production teams to plan well in advance, reducing maintenance costs and downtime.

Using predictive maintenance with non-stationary rotating equipment.

Non-stationary rotating equipment, unlike stationary rotating equipment, doesn't operate at a consistent speed. Depending on the speed at which a machine operates, it will produce different vibrations—a higher operating speed produces higher frequency vibrations.

This makes non-stationary rotating equipment more difficult to monitor. Predictive maintenance solutions have to see past the noise produced by variable speeds, to see the signal indicating a machine fault. AiSight's predictive maintenance solution has two tools to deal with this problem. First, our machine-learning algorithms are capable of comparing vibrations and, therefore, seeing past the noise. Second, our Aion sensor nodes include magnetic field sensors to provide additional perspective on operating speed. The result is a fool-proof solution, capable of detecting faults and predicting breakdowns in non-stationary rotating equipment.

How does predictive maintenance work with so many different machines?

Most machines in modern production facilities use rotating components. As long as something rotates, we can monitor it, predict faults, and prevent unplanned downtime.

In future articles, we'll take a close look at specific machines used in various industries. This series will start with some of the most general types of rotating equipment, found in production facilities everywhere, often as part of more complex machines. This includes motors, pumps, and fans. Because rotating equipment appears almost everywhere in modern production facilities, the possibilities are almost endless.

Using predictive maintenance with motors

Electric motors run most machines in a production line. We find them running pumps, fans, saws, and other rotating equipment either directly driven, belt driven, or driven through a gear box. Electric motors also keep products moving on automatic conveyor belts, and run complex robotics. Whatever the machine, if its motor breaks down, it doesn't work.

AiSight's predictive maintenance solution will prevent that from happening. Our solution provides detailed analysis on bearing faults, identifying problems with the inner chase, outer chase, or ball bearings. It can also identify looseness, imbalances, misalignments, friction, and problems with belt drives and gearboxes connected to the motor. Finally, we can get a look inside motors themselves to identify rotor and stator damage.

Using predictive maintenance with pumps

Pumps keep production flowing. Industries from beverages, to metals, to textiles rely on pumps to get liquids where they need them—whether they're bottling iced tea, cooling a die, or dyeing tees. Not only are pumps important for production, pumps are an important part of waste and HVAC systems. These systems going down can close down an entire plant. And their value is reflected in maintenance costs—our analysis finds that pumps soak up 8% of maintenance costs. It's worth it; keeping pumps running keeps production running.

AiSight's predictive maintenance solution makes that possible by providing early warning of failures in pumps. This includes monitoring faults in the motors running pumps, but also pump-specific problems such as cavitation. AiSight's predictive maintenance solution also detects imbalances and looseness in impellers.

Using predictive maintenance with fans

Whether production calls for heating or cooling, ventilation or extraction, fans are a critical part of production facilities. Individual machines need fans to keep things cool. Convection ovens need fans to distribute heat. HVAC systems use fans to keep facilities comfortable and safe.

We're fans of fans. That's why AiSight's predictive maintenance solution provides effective monitoring and early warning of faults that can stop them from working, including imbalances, looseness, misalignments, and bearing defects.

How does a predictive maintenance solution monitor rotating equipment?

To say that rotating equipment is all AiSight can see is fair—rotating equipment is everywhere. At AiSight, we're serious about machine diagnostics. We want rotating equipment to run as much as possible, for as long as possible. We're leveraging the sensors and algorithms to do so. The specifics of how that works depend on the specific machine. That's why we've launched this series: to look at the machines individually and see how our solution predicts breakdowns months in advance.

To discover how we make unlimited machine uptime a reality, look out for the next article in this series. We'll take a deeper look at motors—at what they do, how they do it, how they breakdown, and how we can prevent it.