Two topics keep coming up in the CIGRE 2024 B5 discussions: network redundancy and time synchronization. Indeed, these are two fundamental design characteristics that ensure the digital substation (DSS) meets the grid’s high expectations for reliability.

This is a summary of my findings related to PRP v. HSR.

In a digital substation, trip signals are sent as GOOSE messages on the LAN networks using Ethernet multicast services. Sample Values are also sent using multicast to data subscribers. Ethernet does not provide guaranteed high availability, as required for critical industrial communications. Thus, the IEC 62439-3 (2021) standard has defined the PRP and HSR protocols to guarantee that these messages and data reach their destination.

PRP (Parallel Redundancy Protocol) and HSR (High-availability Seamless Redundancy) are two industrial redundancy protocols adopted to enhance communication reliability and availability in digital substations, particularly in systems based on IEC 61850. Both support parallel communication to provide zero recovery time in case of a single network failure, but they differ significantly in their implementation.

PRP Network

PRP (Parallel Redundancy Protocol) networks use two completely independent local area networks (LAN A and LAN B) to ensure communication redundancy. A device has two ports (dual attached node) connected to each network to send and receive data simultaneously on both networks. If one network fails, communication continues uninterrupted on the other network. The devices are programmed to ignore the second message once the first message is received.

PRP can be integrated into legacy network architectures since it can work with any topology, including ring, star, or mesh. PRP’s independence from the physical topology makes it a good choice for larger and more distributed networks.

PRP is also compatible with non-redundant devices using a Redundancy Box (RedBox), which translates between PRP and non-redundant networks. This makes PRP suitable for hybrid systems where not all equipment supports full redundancy.

Since PRP uses two fully independent networks, the failure of one network (LAN A or LAN B) is entirely transparent to devices. The system continues operating seamlessly, as data is already transmitted on the other network. Recovery time is zero since both networks are active at the same time. There is no delay in switching over in case of failure.

PRP transmits data over two separate networks, effectively doubling the network traffic, which can be a concern in bandwidth-constrained environments. Each message is transmitted twice, once on each network, which could reduce overall network efficiency if not properly managed. Also, PRP networks require duplicate network switches, thus practically doubling the cost of networking assets.

PRP is best for complex or distributed substation networks where different parts of the system may have different topologies or equipment configurations. It is ideal for situations where seamless integration with non-redundant or legacy systems is needed.

HSR Network

HSR (High-availability Seamless Redundancy) networks use a ring topology where each device has two network interfaces (switching node) to receive and forward messages to the next device in the ring. Data is sent in both directions around the ring, and the first packet to reach the destination is accepted while the duplicate is discarded. HSR relies on a single ring network rather than two separate networks, meaning the devices handle redundancy by forwarding data around the ring. HSR also offers zero recovery time because data is sent simultaneously in both directions around the ring. If one path fails, the data continues to flow in the opposite direction without disruption.

HSR uses a single network ring but still sends data in both directions. However, because it’s confined to a ring, it is generally more efficient in bandwidth usage than PRP. HSR’s ring structure makes the data transmission more efficient for certain configurations, but this advantage diminishes if multiple rings are necessary for larger systems.

HSR requires a ring topology, which may only be suitable for some substation designs. It doesn’t easily integrate with other network topologies without additional devices, and the ring architecture limits the flexibility in network design. If multiple failures occur in the ring, the network could become isolated.

HSR is less scalable because it relies on a ring topology. The number of devices in the ring is limited by network performance and the number of hops, which can increase latency. HSR’s single ring design can make large-scale deployments more complex, particularly if multiple rings are needed to support different parts of a large substation.

HSR  is best for small to medium-sized networks with simpler layouts. If the entire substation can be configured as a ring network, HSR can offer excellent performance and lower hardware costs (since fewer network interfaces may be required).

Summary

PRP is generally better for larger, more complex substations where flexibility, scalability, and integration with legacy systems are essential. It allows the use of multiple topologies, providing greater network design freedom.

HSR is well-suited to small or medium-sized substations where a ring topology can be implemented without significant complexity. In such cases, it offers better bandwidth efficiency and simplicity but is less flexible in larger or distributed networks.

Ultimately, one may want to standardize all transmission and large distribution DSS on PRP as it is considered better in environments requiring high flexibility and scalability.

The evolution of DSS towards software-defined substations that fully utilize software-defined networking (SDN) will undoubtedly bring new ways to guarantee the reliability of data communication. I will discuss SDN performance for DSS in a future article.