Key technologies defining smarter factories: sensors

Sensors can be used in relation to an array of different parameters – from determining temperature and pressure levels in industrial processes through to vibration monitoring of equipment and the current location of important assets. The health of motors can be derived from sensors situated inside the coils, while light sensors can be used for more accurate measurement of the positioning of apparatus, such as the deflection of cutting blades or the orientation of a robot arm.

Image sensors are now being implemented in many different parts of industrial processes – for optical inspection purposes, and even to guide robots using image recognition software. Camera feeds can also be delivered into the central database for machine learning (ML) and monitoring.

The packaging in which sensors are supplied is a key factor to consider. Sensors used across the industrial process are often going to be located in harsh environments, and must operate reliably while exposed to high temperatures, corrosive atmospheres and elevated levels of humidity, as well as coping with shock and vibrational forces. At the same time, they generally need to be compact and support low-power operation.

Temperature sensing

Positive temperature coefficient (PTC) thermistor components produce a rise in resistance when the temperature increases, while negative coefficient ones will show a drop. As well as measuring the local temperature, they can be used as temperature limiters – working in conjunction with a microcontroller unit and control software to trigger an alarm when a preset threshold is surpassed.

These have a higher degree of accuracy and better linearity than thermocouples, though they don’t cover such an extensive temperature range. Temperature sensors are also increasingly integrated directly into the controller chips themselves, in order to provide monitoring of the constituent electronics.

Pressure sensing

A wide range of sensors are needed to encompass all the pressures that can be feasibly experienced in modern industrial processes. Sputtered thin-film strain gauges have proved reliable and accurate under extremely challenging operational conditions – covering ranges from 0 to 100psi right through to 0 to 30,000psi.

Variable capacitance transducers handle lower pressures, typically 0 to 2psi up to 0 to 15psi. These utilise the change in capacitance between two plates as the pressure level alters. Micro electromechanical systems (MEMS) pressure sensors have grown in popularity over the course of the last decade.

These are effectively an extension of the capacitive sensor, using micro-machined capacitor plates that have been built into silicon substrates. While these can only handle lower pressures, they have a considerable advantage – as they can integrate all the supporting signal-conditioning circuitry. This means they can deliver better signal integrity (despite industrial settings being intrinsically noisy), while also saving space, and offering system simplicity plus greater operational longevity.

Determining movement and position 

Sensing apparatus is needed to acquire real-time data on the positioning of moving parts within industrial machinery to ensure repeatability, correct alignment, etc. Of increasing importance is ascertaining the proximity of robot arms in relation to staff, in order to prevent injury.

MEMS-based multi-axis accelerometers and gyroscopes are employed for such purposes, providing details on direction and speed of movement. Now these devices are often being combined in the form of inertia measurement units (IMUs), which can furnish robots with valuable balance and orientation data.

Vibration sensors

The sensing of vibration forces can be valuable in an industrial context. Vibration sensors relying on piezoelectric crystals can generate a current when they vibrate, though MEMS accelerometers are starting to supplant these components, as they offer greater sensitivity and are also far less susceptible to temperature changes (not needing compensation mechanisms as a result).

Vibration sensors are often tuned to the resonant frequency of a particular piece of equipment when it is running normally, so that any deviation can be detected. This often gives early indication of problems within a motor or rotating piece of equipment, meaning that repair work can be undertaken at an early stage.

Light sensors

There are many light sensors on the market. These range from simple infrared (IR) photodiodes or phototransistors that can detect when a beam is broken (as employed in motor encoders, for example), all the way up to high-resolution image sensor arrays for detailed quality inspection tasks. In the coming years more sophisticated imaging systems will become commonplace inside our factories, especially as we see robotics proliferate.

Broadening the scope

The sensing needs of Industry 4.0 won’t just be restricted to production facilities. Through access to high-precision location data, materials can be monitored as they are shipped around the world. This can be handled in a number of ways, often via satellite-based positioning. US GPS, Russian GLONASS and Chinese BeiDou satellite constellations are soon to be joined by the European Galileo network.

Most receiver chips cover all four constellations, providing more-than-adequate choice for equipment developers. These navigation sensors need a wireless link to transport the data to a centralised database. This can be a link to a local wireless network (such as cellular infrastructure) that regularly pings the location of the material when it comes within range of a phone network.

Power issues

Power consumption is a major concern when deploying sensor hardware. Running additional power lines around the factory is not really practical or cost-effective (though it should be noted that the increasing pervasiveness of power-over-Ethernet has helped in this respect). Neither is having to replace hundreds of thousands of batteries each year. Reducing the power budget of sensors and their accompanying communications transceivers can have a positive impact on the overall operating cost of the factory. As a result, there is a drive to eliminate the need to replace batteries.

Energy can be harvested from the surrounding environment to power the sensors. This can be done by using temperature differences, vibrational forces or even radio waves. Thermal energy generators (TEGs) are shrinking in size and increasing in power output. Some can now provide the energy necessary to run a sensor node from a temperature difference of as low as 2°C.

The same piezoelectric sensors that are used for vibration monitoring can also generate power from those vibrations. This move to energy harvesting should help to significantly reduce the cabling for wireless sensor nodes and enable them to be placed inside equipment where replacing a battery is almost impossible.

Conclusion

The ability to monitor machinery, procedures and materials in real time gives manufacturers much greater visibility of numerous activities and advanced warning of any possible problems. The next generation of industrial processes will be able to go beyond real-time monitoring, though. Data supplied by the sensors will be fed into digital simulation models of each particular process or may even be used to model complete factories. This is a topic we will look at in depth later in our blog series.

What will this blog series cover?

1. Key technologies defining smarter factories – connectivity
2. Key technologies defining smarter factories – sensors
3. Key technologies defining smarter factories – the rise of cobots
4. Key technologies defining smarter factories – digital twinning
5. Key technologies defining smarter factories – AI
6. Key technologies defining smarter factories – data security