The topology defines how the substation switches are connected and what happens when one fails. A star is simple and cheap, but its central switch is a single point of failure; a ring tolerates the loss of one link at the cost of more hops and a redundancy protocol; a dual star with PRP loses no frames. The choice depends on the function carried by the network and on the voltage level.
What the topology decides
The topology is the way the substation switches and IEDs are connected. It defines three things that matter to protection and control:
- What happens when a link or a switch fails: how many IEDs are cut off and for how long.
- How many hops a frame makes between two IEDs, and therefore its latency.
- How much it costs: number of switches, ports and fibers.
IEC TR 61850-90-4, the IEC 61850 network engineering guideline, compares topologies, redundancy and time synchronization so the designer can choose on solid criteria. It does not impose one: the choice belongs to the project. The general building blocks of the network are covered in the substation communications network guide.
Star
All IEDs, or the bay switches, connect to a central switch.
- For: it is the simplest, with a single hop between bay switches and the core, and easy to maintain: the failure of a bay switch only affects its own bay.
- Against: the central switch is a single point of failure. If it goes down, the whole substation loses communications.
- Where it is used: medium-voltage substations and industry, where the philosophy accepts losing the network during a repair because trips are hardwired.
Dual star
Two independent stars, LAN A and LAN B, with each IED connected to both through PRP.
- For: the failure of any switch or link in one LAN loses no frame, because the other LAN delivers the copy. It is the usual station bus topology at high and extra-high voltage.
- Against: it doubles switches, power supplies and fibers, and the two networks must be physically separated for the redundancy to be real.
Ring
The switches are connected in a closed chain. To keep frames from circulating forever, a protocol blocks one link in normal service and enables it if another fails: RSTP or MRP. It can also be an HSR ring, where each frame travels the ring in both directions.
- For: it tolerates the loss of a link or a switch with little cabling: two fibers per switch.
- Against: with RSTP or MRP, frames are lost while the network reconfigures. Frames cross several switches, and each hop adds latency. A two-point failure splits the ring into two islands.
- Where it is used: high-voltage station buses when the philosophy accepts reconfiguration; medium-voltage switchgear with HSR.
Ring of rings and ladder
In large substations, a single ring with every switch accumulates too many hops, and one failure affects the whole network. It is split as follows:
- Ring of rings: a main core ring and secondary rings per voltage level or group of bays, each joined to the core by two switches. A failure stays contained within its ring.
- Ladder: two parallel chains of switches joined at several points, like the rungs of a ladder. It gives several alternate paths and pairs well with PRP when each chain is a LAN.
These variants give more availability, but they complicate the latency calculation and the redundancy configuration. They should be documented with a drawing that shows which link the protocol blocks in normal service.
Comparison
| Topology | Failure of one switch | Typical hops | Cost | Typical use |
|---|---|---|---|---|
| Star | If it is the central one, the whole network is lost | Few | Low | Medium voltage, industry |
| PRP dual star | No frame lost | Few | High (two networks) | HV and EHV station bus |
| RSTP or MRP ring | Reconfiguration with a brief loss | Several, depending on position | Medium | HV station bus |
| HSR ring | No frame lost | Several; traffic is doubled | Medium | Process bus, MV switchgear |
| Ring of rings or ladder | The failure stays contained | Several | Medium to high | Large substations |
Latency per hop
Substation switches usually forward a frame only after receiving it completely. Each hop adds that reception time, which depends on the frame size and the link speed: at 100 Mbit/s, a 1500-byte frame takes about 120 µs to be received. If the frame also waits in the queue behind another long one, the latency grows. That is why:
- trip GOOSE and sampled values are tagged with high priority per IEEE 802.1Q;
- links between switches are 1 Gbit/s when there are many hops;
- the number of switches in the path of critical functions is limited.
Station bus versus process bus
The station bus links the IEDs, control system, gateway and clock of the whole substation. Its traffic is moderate (MMS, GOOSE, time) and it accepts a ring, a dual star or combinations of them.
The process bus links merging units, protection IEDs and breakers. It carries sampled values as a constant, heavy stream, and any loss blocks protection functions. That is why it is designed per bay or per protection zone, with PRP or HSR, making sure the traffic of one bay does not flood the others.
- Does the network carry GOOSE trips or sampled values?
- YesDoes the budget allow two independent networks?
- YesDual star with PRP
- NoHSR ring with few switches
- No, only MMS, SCADA and non-critical interlocksIs a brief loss during reconfiguration acceptable?
- YesRing with RSTP or MRP
- NoDual star with PRP
Common mistakes
- A ring with no drawing of the blocked link, which nobody can locate when troubleshooting.
- Switches of the two PRP LANs fed from the same power supply, or with fibers in the same trench.
- Too many switches in a ring, with latencies nobody calculated.
- Mixing RSTP and PRP without a clear criterion on the same ports.
Network design with Tiza Group
Choosing the topology is part of network design and must be done with the timing of each function in view. Our team does it as part of our communications engineering service, remotely, for medium-, high- and extra-high-voltage substations and for industry, for projects worldwide. On site, in Colombia or elsewhere on request, we configure and test the network through our IEC 61850 engineering service. Write to us with the number of bays and the protection philosophy.
Related guides
- Substation network redundancy: RSTP, PRP, HSR or MRP
- IEC 61850 transfer time: classes and measurement
- Substation communications by voltage level: MV, HV, EHV
- Substation IP addressing and VLAN plan
- Substation communications network testing
References
- IEC TR 61850-90-4:2020, Network engineering guidelines.
- IEC 62439-2:2021 (MRP) and IEC 62439-3:2021 (PRP and HSR).
- IEEE 802.1Q: VLANs and priority.
- IEC 61850-5: communication requirements and transfer time classes.
Frequently asked questions
Which is better for IEC 61850, ring or star?
It depends on what the network carries. A ring with RSTP or MRP tolerates the loss of one link, but loses frames while it reconfigures. For GOOSE tripping and sampled values, a dual star with PRP or an HSR ring is preferred, since they have no recovery time.
How many switches can a ring have?
The limit comes from the redundancy protocol and the accumulated latency. MRP, for example, specifies its recovery times for rings of up to 50 devices. In large substations it is more common to split the network into several rings, one per voltage level or group of bays, joined to a core.
Is the process bus topology the same as the station bus topology?
Not necessarily. The process bus carries sampled values with a high traffic volume and strict timing, so it is usually designed per bay or per protection zone, often with HSR or PRP, while the station bus covers the whole substation.