IEC 61850 Station Bus vs Process Bus
Digital-substation architectures split the Ethernet into a station bus and a process bus, and engineers need to know what lives on each and why the separation exists. This page compares the two networks - the services they carry, the reasons they are kept apart, and how devices bridge them - so you can read or design a substation network with confidence.
Station Bus vs Process Bus in one line: The station bus carries IEC 61850 MMS client-server traffic and GOOSE between bay-level devices and SCADA. The process bus carries Sampled Values and GOOSE between primary-plant interfaces (merging units) and protection. They are usually separate networks because the process bus runs high-rate, timing-critical streams that must not compete with routine station traffic.
Two Networks, Two Jobs
In a conventional substation everything shared one network, but the digital substation splits traffic by timing criticality. The station bus is the upper network: it connects bay controllers and protection relays to each other and to the station computer and SCADA gateway. Its traffic is MMS reports and control plus station-level GOOSE. It behaves like a normal, if carefully engineered, industrial Ethernet.
The process bus is the lower network, sitting between the primary plant and the protection and control devices. Its defining traffic is Sampled Values - continuous digitized current and voltage from merging units - plus fast GOOSE for trips and status to the switchgear interface. Because SV is a steady high-rate stream that protection depends on cycle by cycle, the process bus is engineered for very low loss and tight time synchronization, and it is kept separate so routine station traffic cannot disturb it.
How Devices Span the Two
Many devices touch both buses. A protection relay subscribes to Sampled Values and GOOSE on the process bus to make its trip decision, and it publishes reports and responds to control over MMS on the station bus. The merging unit lives on the process-bus edge, turning analog secondaries into an SV stream. The SCADA or cloud gateway usually connects only to the station bus, where it speaks MMS.
The separation is also a resilience and security boundary. Keeping Sampled Values and process-bus GOOSE on their own segment, with VLAN and priority tagging, protects protection-grade timing and limits blast radius - a principle that connects to broader SCADA network segmentation and subnetting practice. Both buses are facets of the same IEC 61850 standard.
Side by Side: The Two Buses at a Glance
The quickest way to hold the architecture in your head is a straight comparison. The station bus is a routable TCP/IP network doing supervisory work; the process bus is a Layer 2 multicast network doing measurement and tripping work. Everything else - the hardware choices, the time sync, the redundancy scheme - follows from that split.
| Aspect | Station bus | Process bus |
|---|---|---|
| Main traffic | MMS reports and control, station-level GOOSE | Sampled Values, trip and status GOOSE |
| Transport | TCP/IP, routable | Layer 2 multicast, not routable |
| Timing need | Event ordering across devices | Sample-accurate alignment |
| Typical devices | Relays, bay controllers, station computer, gateways | Merging units, relays, switchgear interfaces |
| Time sync | NTP or SNTP class | PTP with hardware timestamping support |
Note what the table implies about failure tolerance. A brief pause on the station bus delays a report; the same pause on the process bus punches a hole in the current and voltage stream a relay is deciding on. That asymmetry is why the two networks are engineered to different standards even when they are built from similar parts.
Time Sync: Event Ordering vs Sample Alignment
The station bus needs clocks good enough to order events across devices - when a breaker opened relative to when a protection element picked up - and NTP-class synchronization over the network is the usual answer; see NTP time synchronization for how that works in a SCADA context. The process bus is a different animal: Sampled Values from different merging units must be aligned sample by sample before a relay can compare or sum them, so the network distributes time with PTP (IEEE 1588), with hardware timestamping support in the switches and a power-utility profile defined for exactly this purpose.
Loss of synchronization means different things on each bus. On the station bus it degrades timestamps and makes post-event analysis harder. On the process bus it is a protection-affecting condition: merging units flag their streams as unsynchronized and relays react according to their configuration - blocking some functions, falling back on others. What a given relay does is device-specific, so the behavior belongs in the protection design review, not in assumptions.
Redundancy: Why PRP and HSR Show Up Here
Conventional Ethernet redundancy recovers by reconverging - a spanning-tree topology change, a ring healing - and reconvergence takes time during which frames are lost. The process bus cannot tolerate that gap, which is why digital substations lean on the IEC 62439-3 schemes: PRP, which duplicates every frame across two fully independent networks, and HSR, which circulates duplicates both ways around a ring. In both, the receiver takes the first copy of each frame and discards the second, so a single network failure costs zero frames.
PRP doubles the network hardware; HSR saves hardware but adds ring design constraints. Which scheme to use, and whether the station bus gets the same treatment or lives with conventional resilience, is a site engineering decision driven by the protection scheme's tolerance for loss - there is no universal answer, and utility standards differ on it.
Retrofits: Adding a Process Bus to a Working Station
Most existing IEC 61850 stations are station-bus-only, with copper CT and VT secondaries landed directly on the relays. A process bus arrives incrementally: one bay gets merging units, its relays subscribe to Sampled Values instead of - or alongside - the copper inputs, and the rest of the station carries on unchanged. The station bus barely notices, since MMS reporting and control continue as before, which is what makes bay-at-a-time migration practical.
The part that changes character is testing. With copper, a technician injects current at the relay terminals; with a process bus, test signals become published SV streams and simulation flags in the frames, and isolation moves from physical test switches to subscription configuration. That shift needs planning, updated procedures, and protection engineers who have signed off on the approach - it is the single biggest cultural change in the migration. Where fast data must also leave the station, the routable profiles of IEC 61850-90-5 pick up where the local buses stop.
Frequently Asked Questions
What traffic runs on the process bus?
Primarily Sampled Values - continuous digitized current and voltage from merging units - plus fast GOOSE to the switchgear interface. It is a high-rate, timing-critical network that protection relays depend on cycle by cycle.
Why separate the station bus and process bus?
Because process-bus traffic (Sampled Values and protection GOOSE) is high-rate and timing-critical, it must not compete with routine MMS reporting and control on the station bus. Separation protects protection-grade timing and provides a security boundary.
Does every substation have a process bus?
No. Many IEC 61850 substations use only a station bus with conventional copper secondaries wired to the relays. The process bus appears in digital substations that use merging units and Sampled Values instead of copper analog wiring.
Can the station bus and process bus share the same physical switches?
It is technically possible to separate them with VLANs on shared hardware, and some small-station designs do. Common practice is physical separation, because it makes the protection-grade network's behavior independent of station-traffic mistakes. Which way to go is a site engineering standard, not a protocol requirement.
Does SCADA ever connect to the process bus?
Normally no. Supervisory systems and gateways sit on the station bus and receive their data over MMS reporting; the process bus serves protection and control devices that consume Sampled Values directly. Keeping supervisory clients off the process bus is part of the point of the separation.
Automation services
Need help turning this into a working system?
Merobix integrates SCADA, programs Allen-Bradley and Siemens PLCs, and designs and fabricates industrial control panels.
Meeting requests are reviewed before confirmation.