Standards and requirements context
The standards below frame the design. None of them replaces a site-specific risk assessment.
- IEC 62591 covers WirelessHART, a wireless communication profile for process automation field devices.
- ISA100.11a (also published as IEC 62734) is the competing industrial wireless standard for process control and related applications.
- IEEE 802.15.4 defines the low-rate, low-power radio layer that both industrial mesh standards are built on.
- IEC 61511 governs functional safety for safety instrumented systems in the process industry. Wireless monitoring is normally kept outside safety functions unless a documented justification exists.
- IEC 60079 (the explosive atmospheres series) governs equipment selection for classified areas. Every transmitter, antenna, power module and gateway must be certified for the area in which it is installed.
- IEC 62443 is the series on security for industrial automation and control systems. It applies to the gateway, its network connection and the device join process.
The NIST guide to industrial wireless deployments and the IEEE Industrial Electronics Magazine paper on guidelines and the deployment life cycle (2018) both treat wireless as a life-cycle activity. Survey, design, installation, commissioning and ongoing monitoring all count. A one-time installation does not.
Choosing the technology for the job
The Emerson planning material and the iFactory commentary criticise the same habit: picking one technology for everything. Sometimes Wi-Fi gets chosen because IT already runs it. Sometimes a low-power wide-area IoT network gets pushed as the universal answer. The usual outcome is dead batteries, lost data and dead zones.
| Criterion | Self-organising mesh (WirelessHART, ISA100.11a) | Wi-Fi | Long-range low-power (LoRaWAN-type) |
|---|---|---|---|
| Typical role | Process variable monitoring from many field devices | Mobile workers, video, high-bandwidth tasks | Sparse, slow-changing, widely scattered points |
| Topology | Mesh or star-mesh | Star (access point) | Star |
| Power profile | Designed for battery-powered field devices | Poor fit for long battery life | Very low power, low data rate |
| Industrial process alignment | Purpose-built, with standard device profiles | General-purpose | General-purpose |
| Hazardous-area devices | Widely offered | Limited | Varies by vendor |
| Main risk on a brownfield site | Poor device placement and gateway planning | Coverage gaps and contention in steel-dense areas | Latency and data-rate limits for process data |
For retrofit process monitoring in classified areas, the mesh standards are usually the first candidates. The other technologies earn their place where their strengths match the job. Wi-Fi for mobile operators is a good example.
Treat any vendor capacity limit as a ceiling for one gateway. A design that sits right at the ceiling has no room for later additions, or for rerouting around a failed node.
Topology design
Star, mesh, and star-mesh
In a star network, every device talks directly to the gateway or access point. It is simple and it gives low latency. The catch is that any obstruction between a device and the gateway becomes a single point of failure. A mesh lets each device relay for its neighbours, so there are alternative paths when a link degrades. A star-mesh uses mains-powered routers as the backbone and battery devices at the edges.
On a brownfield site, mesh is usually the better bet, because the radio environment is hard to predict. Tanks, columns, pipe racks and vessels keep changing which paths are usable. A mesh gives the network several ways around them.
Design on paper first
The Emerson PlantWeb University course on self-organising networks lays out a sequence that holds up in practice:
- Scope the project.
- Envision device locations.
- Network the devices on paper.
- Integrate the gateway.
- Install the network.
- Fortify the network.
Most of the savings come from doing the paper design before any hardware arrives. Emerson's planning article also mentions software that takes an uploaded plot plan and checks a proposed design. Those tools help, but they only know what the drawing shows, and brownfield drawings are often out of date. Someone still has to walk the site and verify obstructions, particularly recent additions such as new vessels or scaffolding.
Placement rules that matter
- Give every device multiple neighbouring devices or routers it can reach, so no device depends on a single link. The planning sources describe this principle, and the exact neighbour count should follow the vendor guidance for your system.
- Place the gateway centrally and high, with a clear view of the plant, rather than in the most convenient equipment room.
- Put routers where device clusters are dense, not only where power happens to be easy to reach.
- Keep antennas clear of large metal surfaces, and use the orientation the vendor specifies.
- Add routers for future expansion zones as well as the current scope.
Reliability design
Radio resilience
The industrial mesh standards use channel hopping and time-synchronised communication, which lets them shrug off interference that would take down a fixed-channel system. That is not a licence to skip the homework. The NIST guide stresses that the radio environment has to be understood before commissioning. Survey the existing radio users (Wi-Fi, radios, other wireless instruments) and plan how the new network will coexist with them. A spectrum survey skipped ahead of a Wi-Fi expansion or a contractor's temporary link is a common cause of intermittent faults later on.
Path diversity and gateway redundancy
Design so that losing any one router or link does not cost you any device. Prove it by switching off a router during commissioning and watching what the network does. The gateway is the real single point of failure. Where the monitoring matters operationally, consider a redundant gateway arrangement if the system supports one, and give the gateway a resilient backhaul to the control system or historian.
Integration and data quality
The gateway usually passes data to the DCS, historian or asset management system over standard protocols such as Modbus or OPC. Agree the data model early:
- Define how the DCS flags stale or lost wireless data. A frozen last-good value shown as live is dangerous.
- Set update rates according to process dynamics, not as a default.
- Keep device diagnostics visible to maintenance as well as operations.
Security
Treat the wireless network as part of the control system's attack surface. Use the device join credentials the standards provide, restrict who can provision devices, segregate the gateway from the business network, and apply the IEC 62443 approach of zones and conduits. An unmanaged gateway on a flat network wipes out the benefit of the encryption built into the radio layer.
Power management
Battery life is the main running cost of a wireless network, and most of it is settled at design time.
What drives battery consumption
- Update rate. Faster publishing uses more energy. Many monitoring points do not need fast updates, and slowing them stretches the time between battery changes.
- Routing duty. A device relaying traffic for others uses more energy than a leaf device. If a battery device sits on a busy path, add a powered router to take over the relaying.
- Temperature. Hot or cold locations cut battery performance. Check the vendor's conditions against the real location.
- Join and diagnostic traffic. Repeated rejoining after poor placement drains batteries fast.
Maintenance planning
Keep a register of every device, with its installation date, power module type and last replacement. Use the device-reported battery status in the asset system to schedule changes in batches, so they are never emergency call-outs. Replace power modules only with types certified for the area, and follow the site permit rules. In classified areas the module change is itself a hazardous-area task and needs the same permit discipline as any other work there. Where mains or solar supply is available, powered routers and harvesting-assisted devices cut the long-term workload. Each addition still needs its own area certification and installation check.
Illustrative scenario
This scenario is constructed for illustration and is not drawn from a listed source.
A tank farm next to a processing area needs bearing temperature and discharge pressure monitoring on the transfer pumps, plus level confirmation on several tanks. The nearest marshalling cabinet sits a long way off, across a congested pipe rack. Cable is impractical.
The design team works through the sequence above:
- Scope. All points are monitoring only. None feeds a trip or safety function, so the IEC 61511 boundary holds.
- Survey. Walk the site, mark the tanks, bunds and rack steel on the current plot plan, and check which radio systems already operate there.
- Gateway. Mount it at height on an existing structure in the processing area, with a clear view over the tank farm and a connection to the control network.
- Routers. The tanks block direct paths, so add powered routers on lighting masts at the farm edge. Battery transmitters then have several alternative paths.
- Update rates. Bearing temperature and level move slowly, so they get relaxed update rates. Pressure on the transfer line is published faster for operations.
- Commissioning test. Switch off one router. Confirm that every device stays connected and that the DCS flags data quality correctly.
When a transmitter has to go onto a process tapping, the work follows the normal intervention procedure. Isolate the tapping, depressurise, verify zero energy, apply lock-out/tag-out, and work under a hot-work or cold-work permit with gas detection active. Vent safely before opening. Transmitters mounted on existing thermowells or in non-intrusive surface positions avoid opening the process at all, which suits a brownfield retrofit well.
Practical checklist
- Confirm each monitoring point is outside any safety instrumented function, or document the justification under the IEC 61511 lifecycle.
- Verify that each transmitter, router, antenna, power module and gateway is certified for its installed area classification.
- Update the plot plan from a physical walk-down before any design software is used.
- Survey existing radio systems and record coexistence decisions.
- Give every device multiple reachable neighbours, and test by removing a router.
- Keep the planned device count below the vendor's gateway limit, with room for growth.
- Decide gateway location, redundancy and backhaul before ordering.
- Set update rates per variable, not as a global default.
- Define stale-data handling and diagnostics display in the DCS.
- Apply IEC 62443 zoning, join-credential control and access restrictions to the gateway.
- Set up a device register and battery replacement schedule, and tie it to the permit system.
- Re-survey after major plant modifications.
Conclusion
Wireless instrumentation is a practical way to close monitoring gaps on brownfield sites, as long as it is engineered like any other infrastructure. Do the paper design first, pick the technology for the application, and build in path diversity and a resilient gateway. Plan battery replacement as managed maintenance, and keep security and hazardous-area certification in scope from day one.
As a next step, pick one unit with several known monitoring gaps. Run a walk-down and radio survey, then pilot a small mesh with a spare-router failure test before extending it to the whole site.