Sensors can be used to detect, for example, location, light level, proximity, movement, falling, magnetic, leaks, force/load/strain, flow, temperature, humidity, sound, chemical/gas, air pressure and smoke.
When choosing sensors you need to think how data will end up at your server or a cloud server because the sensor needs to be compatible with your chosen system. Here we cover the main types of system and their key strengths. There’s no ideal generic system and your choice will depend on your existing setup and requirements.
Programmable Logic Controller (PLC) Based
PLCs are what have been used since about 1968 to provide for control and automation in factories. They are well known, reliable, have long life and form what became Industry 3.0. They are built for industry being shock and vibration resistant, electrical noise resistant, corrosion resistant and work across a wide ambient temperature range. A PLC is a module that plugs into a backplane together with a power supply and multiple input/output modules. The input/output modules are wired to sensors and actuators attached to and within physical machines. The units are linked by wire using a variety of protocols over Fieldbus or, more recently EtherCAT, an extension of Ethernet for industrial automation. The PLC itself is programmed using ‘ladder logic’.
Legacy PLCs aren’t ideally suited to Industry 4.0 and IoT for several reasons. Industry 4.0 needs to collect data from a lot more sensors and requires the addition of a lot more PLCs. The extra data load and typical pyramid implementation where PLCs report upwards means that the system capabilities can become overwhelmed. The conventional poll/response protocols also aren’t ideally suited to IoT which instead tends to use publish/subscribe models that only do work and consume resources when there’s new data.
PLCs are evolving. PLC manufacturers are starting to offer systems that replace the PLC with multiple smaller PLCs. Look for so called nano PLCs, ultra-compact PLCs or micro PLCs. Also look at Programmable Automation Controllers (PAC) that work at a higher level than PLCs and are architecturally similar to a PLC combined with a PC. Keep an eye on new industry standardisations, products and frameworks such as OPC Unified Architecture (UA) and DDS that allow better interoperability. For companies, expecially SMEs, that can’t afford to replace existing PLC-based systems with newer systems, there are also solutions to retrofit and overlay IoT.
These range from Raspberry Pi and Arduino boards intended for learning and hobyists up to proprietary IoT boards, gateways and black boxes. They usually require programming in c, Python, Arduino or a scripting language. Implementation usually also requires soldering, finding cases and other low level problem solving.
Bluetooth LE Based
Bluetooth LE Sensors, sometimes called beacons or tags, advertise their id and sensor information that can be picked up by other Bluetooth LE devices. These include smartphones, microcontrollers or WiFi/Ethernet gateways that send data to your server or the cloud. Bluetooth LE sensors can be powered by batteries that last years. This makes them suitable for hard to reach places or for putting on mobile things or people to track position.
The design of Bluetooth LE is such that it’s electrically noise tolerant and suitable for use on the factory floor. A new standard Bluetooth mesh protocol allows devices to communicate with each other across a site to relay data to overcome the typical 50m to 100m typical range.
Devices are inexpensive and if used with Bluetooth gateways the only effort revolves around setup rather than programming. There’s no soldering or hardware interfacing.
Bluetoooth isn’t as deep sensing as PLAs and can’t detect flow or strain and can’t easily activate anything. It can detects ambient things such as position, movement, temperature, humidity, air pressure, light level, open/closed via magnetic hall effect, proximity (PIR), proximity (infrared cm range), fall, smoke, natural gas and water leak.
Smart Sensors and Smart Devices
There’s also a increase in the availability of ‘smart’ IoT devices and sensors that include connectivity. Some microcontrollers are starting to include NB-IoT or LoRa. Some sensors are also starting to include LPWAN such as LoRa. These ‘smart’ devices can be used on their own or in conjunction with previously mentioned systems.
PLCs are good for deep sensing of things such as flow and load/strain and actuation/activation but unsuitable for sensing mobile things. Bluetooth is the opposite in that it’s good for asset tracking and sensing ambient things. It can also detect quantities when attached to things but unlike PLCs doesn’t become part of the machinery. Microcontrollers are the most flexible but that flexibility comes with the compromise of significant practical set up and programming.
The choice depends on what you have already, what you need to sense and activate, existing skills and your budget. In some cases you will need a hybrid solution.