Are We Ready for Wireless Meshes?

Traditionally, wireless mesh networks have a bad reputation in industrial environments. It’s easy to understand why; radio frequency (RF) is the stuff of black magic – it behaves differently in different environments, it changes in response to changes in humidity, it doesn’t like being around big lumps of metal, or lots of water-filled containers, or humans – especially if they move around a lot. All of these characteristics conspire to make an RF connection the exact opposite of a bit of wire in terms of stability of signal. The net effect is that transmissions fade in and out with time.

Why this should surprise us is a mystery – we are all familiar with these effects in our home wifi networks, when out and about on our mobile phones, even when watching TV, so why do we expect RF systems to behave differently in our industrial environments?

This is certainly not the case with industrial data, where missing a sensor event could have very serious consequences for the cost of a manufacturing run, the safe operation of a gas pipeline or the energy consumed in heating or cooling a building. Because most industrial data is critical, the transmission infrastructure which carries it has to perform reliably 24/7 irrespective of the environmental conditions or the movement of people or machinery. Sadly this has not been the experience of many when deploying meshing radio systems in industrial environments. However, the case for deploying these systems, in terms of reducing the cost of installation of sensor networks, is overwhelming, so we keep on trying.

Things are changing

It looks like things are now changing. The new standard of IEEE802.15.4e adds two features to the formation and running of mesh networks that transform their reliability when being used in difficult environments – namely the addition of time synchronisation across network nodes, and the ability to channel hop between available channels within the frequency band. These features provide two very important benefits, which are set to change the conversation about deploying meshing networks in industry.

The first benefit is around the power requirements for intermediate nodes. In old meshing systems, there was no constraint on when any of the nodes could send data. This implicitly means that any node acting as a forwarding point for transmissions across the network (which is normally most of them) has to remain permanently powered to ensure that they are active when any transmission occurs. This need for permanent power, in turn precludes the use of batteries for the power source as they drain too quickly. Adding time synchronisation means that now we do know when any transmissions are going to take place, and the whole network, including the intermediate nodes, can now enter a sleep state between these times, allowing the whole network to run, depending upon the sampling period, for many years off of small and inexpensive batteries.

Of course, we still need to provide power to the sensor circuits, and so could power intermediate nodes fairly easily in many applications, but when time synchronisation is coupled with the second benefit, the ability to channel hop, things get more interesting. In traditional systems, like home wifi networks, it is only necessary to make a choice about which channel the system will use when the system is installed. We pick what seems to be the clearest channel, but over the course of days, weeks and months, this will fade in and out due to all the reasons discussed earlier. In a time-synchronised, channel-hopping system, the network can look at the available channels and select the best available each time the network wakes up, essentially overcoming the changes in signal quality in any individual channel as, for all practical purposes, there will always be at least one ‘good’ channel free.

B+B SmartWorx recently released a wireless meshing sensor network based on IEEE802.15.4e because it believes that, finally, the technology is fit for purpose, and it is now finding that it can get 99.999% transmission availability in real-world industrial applications. Basically it is now as good as wire, but brings with it the benefits of reduced installation cost benefits that are associated with wireless systems.

The largest single cost element in deploying an asset monitoring system is normally the cost of wiring sensors back to instrumentation cabinets – studies show that this could be reduced by up to 60% by using wireless technologies.