IoT-enabled sensors ready to take on industry 4.0

When it comes to the predictive maintenance and analytics techniques that large industry 4.0 players are constantly developing, the more sensors they have access to, and the better. That’s because with algorithms and machines, more data = more accurate results and predictions. Small, low-power IoT enabled sensors are one of the fundamental components to these emerging technologies, let’s take a look at some of them, how they work and how it all comes together into an industry 4.0 ready ecosystem.

IoT Sensor Technologies 

These small devices employ a wide-range of different sensing methods like: Ultrasonic Infrared, acoustic, electromagnetic, vibration, humidity, chemical, photovoltaic, proximity, pressure, and many more. Let’s take a look at some examples of how these IoT sensors work in predictive maintenance.

Vibration analysis

By far the most widely employed method for predictive maintenance in factories, and even by humans in everyday-life. Cars are an excellent example of an application of vibration analysis to diagnose issues like bad suspensions, worn belts and damaged gears in transmissions. Any equipment with moving parts generates a distinct vibration and signature. By harnessing large data sets on these signatures, and running them though complex learning algorithms, patterns of failure and component degradation start to surface and enable predictive maintenance.

Example: Ball-Screws. These screws are deployed in any machine that requires translation of rotational movement into linear. They are very efficient due to the fact that they are near friction-less. Nonetheless the bearings can degrade overtime causing quality issues in production. By analyzing its vibration frequency patterns and other factors, one can predict the remaining useful life-expectancy of the ball-screws, accurate real-world predictive analytics shown in the chart above.

Infrared Thermography

For predictive maintenance of moving parts, there’s no better source of information than infrared thermography. By using infrared spot sensors and cameras, one can get an accurate reading of the temperatures of components. This can detect temperature anomalies caused by things like poor lubrication causing increased friction, degraded components, faulty electrical connections, stress points, thermal leakage from insulated systems and so on.

Example: With infrared you could detect overheating components and diagnose imminent component failure. The above picture shows how a faulty bearing has caused the entire motor to overheat.

Applied to factory scale = sensor networks

Now apply this with hundreds of IoT-enabled sensors strategically placed in a smart factory, and we have what’s called a sensor network. Considering the placement of these sensors, most of the time the networks are wireless. Zigbee and Bluetooth LE are examples of wireless technologies currently employed in IoT sensors due to their low-power consumption.

This sets the stage for the next crucial piece of technology to take over, cloud computing.

Processing the data

The massive amounts of data and computationally exhaustive simulations/algorithms make cloud computing a nice fit for turning all the unreadable information into a usable state for real-world applications and re-configurations. Here processing can take many forms, from simple sorting algorithms to advanced carbon-copy virtual simulations of actual hardware being run with the sensor data.

Applied predictive maintenance

Now that we have narrowed down root causes, encroaching component failures and degradation patterns, we can start to apply predictive maintenance to the systems. While simple enough to do, the fact remains that human interactions with systems introduce unknowable variables into a system that could skew or even ruin accurate predictive models and simulations. That’s where actuators come in.
there’s the sensing mechanisms, then there’s the mechanisms that reconfigure and change aspects of the physical machine. These emerging systems with sensors, actuators and virtualized software controls are known as CPS (Cyber physical Systems). When these systems become more advanced, things like self-maintenance and self-sufficiency are not out of the realm of possibility.

End goal: Self-sufficient smart factories

With machines now able to handle their own analysis, maintenance, and configuration, self-sufficient smart-factories can be born. By taking all the human elements out of the equation, predictive analytics can become far more accurate introducing a new level of production and efficiency unavailable to the industry as it currently stands. Automation will keep reaching new heights, to the point were humans may no longer understand how the machines they created work and will be unable to service the machines ourselves any longer (like any of us would be working at that point anyway).