Process Bus Architecture: Benefits, Risks and Design Tips

Walk into any recent 400 kV or 132 kV substation project in the UK and you will notice something missing: rows of copper multicore cables that used to run from switchyard CTs and VTs to the relay room. On a site we visited in the north of England last year, the secondary wiring had been replaced almost entirely by fibre. That shift is process bus at work, and it is quietly reshaping how we approach substation automation. It is not simple, and it is not free of risk, but our team believes the complexity is worth it on the right projects. Here is our honest assessment.

What Is Process Bus in IEC 61850?

In simple terms, the process bus is the communication network that connects primary plant measurement and control devices — merging units, CTs, VTs, switchgear — to the protection and control IEDs. It sits below the station bus, which handles inter-IED communication and SCADA links. The process bus is standardised in IEC 61850-9-2 for Sampled Values and IEC 61850-8-1 for GOOSE messaging, which is why the architecture is vendor-neutral in principle, even if practice is messier.

Station Bus vs Process Bus: Key Differences

The station bus carries abstract data models and control commands between IEDs, gateways and the control centre. It operates at higher levels of the hierarchy and is typically Ethernet-based with GOOSE and MMS traffic. The process bus, by contrast, sits at the interface with primary equipment. It carries time-critical, high-volume data — sampled current and voltage waveforms at 80 samples per cycle or more — and trip/block signals via GOOSE. Latency requirements are far tighter, and a failure has immediate protection consequences rather than just SCADA visibility loss.

Sampled Values and GOOSE Messaging Explained

Sampled Values (SV) replace analogue copper from instrument transformers. A merging unit digitises the CT/VT outputs and publishes them as multicast Ethernet frames per IEC 61850-9-2. GOOSE messaging replaces hard-wired binary signals — trip commands, interlocking, breaker status — with publisher-subscriber Ethernet frames defined in IEC 61850-8-1. Together they remove the need for thousands of copper cores, but they also move the reliability question from “will the wire break?” to “will the network deliver on time?”

The Benefits We See in UK Digital Substations

National Grid and SSEN have both run digital substation trials that demonstrate the practical gains. National Grid’s work at substations such as those in its digital substation programme showed significant reductions in secondary cabling and commissioning time. SSEN’s trials in its distribution network reached similar conclusions at 132 kV. The pattern we see across these projects is consistent.

Cost and Space Savings from Reduced Copper

Copper multicore is expensive, heavy and slow to install. On a typical 400 kV bay, replacing hard-wired CT/VT circuits with fibre can cut secondary cabling by 70–80%. That translates into smaller relay rooms, less cable trenching, fewer termination errors and shorter outage windows. On new-build sites, the civil and containment savings alone can be substantial.

Improved Testing and Commissioning

With SV and GOOSE, we can inject test signals digitally at the merging unit or via a test set that emulates the process bus. That means no physical disconnection of CT circuits, no risk of open-circuiting a live CT, and repeatable test scenarios that can be scripted. Commissioning teams we work with report faster FAT and SAT cycles, and fewer wiring-related defects at energisation.

Flexibility for Future Smart Grid Upgrades

Because the process bus is data-driven, adding new functions — power quality monitoring, adaptive protection, wide-area schemes — becomes a software and configuration exercise rather than a rewiring project. That flexibility is central to smart grid integration, where protection settings and monitoring needs will evolve over the asset’s 40-year life.

Risks and Challenges of Process Bus Architecture

We would be doing you a disservice if we pretended process bus is straightforward. The risks are real, and they need to be engineered out at design stage, not discovered at commissioning.

Network Reliability and Latency Concerns

Protection now depends on Ethernet switches. A misconfigured VLAN, a spanning-tree loop or a saturated link can delay or drop GOOSE and SV frames. Latency budgets for GOOSE trip signals are typically in the low single-digit milliseconds, and SV streams consume significant bandwidth. Redundancy protocols such as PRP and HSR are essential, and so is careful traffic engineering.

Cybersecurity Threats in Digital Substations

Every Ethernet port is a potential attack surface. Process bus networks must be segmented, monitored and governed by IEC 62443-aligned policies. Unauthorised GOOSE or SV injection could cause maloperation, so access control, anomaly detection and strict change management are non-negotiable.

Interoperability Between Vendors

IEC 61850 was designed for interoperability, but the reality of mixing ABB, Siemens and GE devices on one process bus still demands rigorous Factory Acceptance Testing and, ideally, System Integration Testing. SCL file handling, SV configuration options and GOOSE performance vary between implementations. Plan for it early.

Design Tips for Process Bus Projects

These are the lessons we apply on every project, drawn from UK deployments and larger offshore programmes.

Network Topology and Redundancy

Keep the process bus topology simple and deterministic. Our default recommendations:

• Use PRP for redundancy where parallel independent networks are practical; use HSR for ring-based architectures where cabling is constrained.
• Separate SV and GOOSE traffic onto dedicated VLANs, and never share a process bus with station bus traffic.
• Size switch buffers and bandwidth for worst-case SV subscription counts, not average.
• Document the network design in the SCL file from day one — it becomes the single source of truth.

Time Synchronization with IEEE 1588

SV and GOOSE depend on accurate time. IEEE 1588 Precision Time Protocol is the standard choice, with grandmaster clocks and boundary clocks distributed across the substation. We recommend redundant grandmasters, GPS holdover, and monitoring of time error at each merging unit. Without solid PTP, sampled values are meaningless.

Testing and Validation Best Practices

Build a full process bus test rig before site works. Validate GOOSE performance under fault conditions, test PRP/HSR failover, and confirm PTP accuracy under load. Lessons from large offshore programmes such as Dogger Bank underline the value of end-to-end integration testing — the cost of finding a network defect offshore is orders of magnitude higher than in a lab.

Our Verdict: Is Process Bus Worth It?

For new-build digital substations at 132 kV and above, and for major refurbishments where secondary systems are being replaced wholesale, yes — the benefits generally outweigh the risks. For small retrofits of ageing 33 kV or 11 kV sites, the business case is weaker, and conventional hard-wired schemes may still be the pragmatic choice. The deciding factors are scale, the availability of skilled commissioning resource, and whether the owner intends to adopt smart grid functions over the asset life.

FAQ

Is process bus mandatory under IEC 61850?

No. IEC 61850 supports both conventional hard-wired interfaces and process bus architectures. Process bus is an option defined in IEC 61850-9-2 and 8-1, and adoption depends on project requirements and owner standards.

Can we mix ABB and Siemens IEDs on one process bus?

In principle yes, because the standards are open. In practice, interoperability requires careful SCL engineering and thorough FAT/SAT testing. We always recommend a system integration test before site deployment.

What redundancy scheme should we use — PRP or HSR?

PRP suits architectures with parallel independent networks and is simpler to troubleshoot. HSR suits ring topologies and reduces cabling. The choice depends on substation layout and switchgear arrangement.

How accurate must IEEE 1588 time synchronisation be?

For protection-grade Sampled Values, we target sub-microsecond accuracy at the merging unit. Redundant grandmasters and continuous monitoring are essential to maintain that in service.

Does process bus meet UK grid code requirements?

Yes, when designed correctly. National Grid and SSEN trials have demonstrated compliance, but the design must account for latency, redundancy and cybersecurity requirements set out in the relevant grid codes and IEC 62443.

Process bus is not a silver bullet. It is a powerful enabler for UK smart grids, provided the network architecture, time synchronisation and testing regime are engineered properly from the outset. Get those right, and the copper savings, flexibility and future-proofing are genuinely worth the complexity.