Sampled Values (SV)

SV operates on a strict publisher-subscriber model where a Merging Unit (MU) continuously multicasts digitized waveforms to multiple subscribing IEDs simultaneously. Unlike polling-based protocols, SV pushes data at a fixed, high sampling rate—typically 80 samples per cycle (4 kHz for 50 Hz systems) or 256 samples per cycle (12.8 kHz for 60 Hz systems). Each Ethernet frame contains a configurable number of Application Service Data Units (ASDUs), allowing engineers to balance bandwidth consumption against latency. The protocol uses VLAN tagging and multicast MAC addressing to segregate traffic, ensuring that protection-class messages are prioritized over monitoring traffic via IEEE 802.1Q priority tagging.

SV is fundamentally dependent on sub-microsecond time synchronization across the substation. The protocol relies on IEEE 1588v2 Precision Time Protocol (PTP) with the Power Utility Automation profile (IEC 61850-9-3) to timestamp each sample at the source. A grandmaster clock distributes time, and boundary clocks within Ethernet switches correct for network delay. The smpSynch flag in the SV frame indicates synchronization health:

• TRUE: The MU is locked to a valid time source with accuracy better than ±1 µs.
• FALSE: The MU has lost synchronization; subscribing IEDs must reject these samples for protection purposes. This allows protection relays to align current samples from different bays for differential protection schemes without dedicated copper wiring.

Every SV frame includes a quality attribute for each analog channel, encoded as a bit-string per IEC 61850-7-3. These bits communicate the trustworthiness of the measurement:

• validity: Indicates 'good', 'invalid', 'questionable', or 'overflow'.
• source: Specifies whether the value is from a live measurement, a substituted manual entry, or a default value.
• test: Marks data generated during maintenance testing, allowing subscribing IEDs to block protection operations.
• operatorBlocked: Signals that an operator has intentionally blocked the value. Subscribers use these bits to make autonomous decisions—a protection relay will immediately block a differential trip if the quality of a remote current sample degrades to 'invalid', preventing nuisance operations.

Deploying SV requires careful process bus network design. A single MU publishing 8 channels (4 currents, 4 voltages) at 80 samples/cycle generates a constant ~5-7 Mbps stream. At 256 samples/cycle, this increases to ~20 Mbps. Key design constraints include:

• Deterministic latency: Switches must support cut-through or low-jitter store-and-forward modes.
• Zero packet loss: Protection applications cannot tolerate lost frames; Parallel Redundancy Protocol (PRP) or High-availability Seamless Redundancy (HSR) are often mandated.
• Traffic segregation: SV traffic is mapped to the highest priority VLAN (Priority Code Point 6 or 7) to ensure it is never queued behind lower-priority SCADA polling traffic. Network bandwidth calculations must account for the aggregate traffic from all MUs on a process bus segment.

SV enables advanced virtual injection testing without physical test sets. A test device can publish SV streams with the test bit set to TRUE in the quality field. Subscribing IEDs configured in 'test mode' will process these injected values for protection scheme validation while the live process bus remains operational. This supports:

• Secondary injection testing: Simulating fault waveforms directly into protection relays over Ethernet.
• Hardware-in-the-Loop (HIL): Connecting physical IEDs to real-time simulators that publish SV streams representing complex grid scenarios.
• Commissioning: Verifying the complete signal chain from simulated primary equipment through the merging unit logic to the protection trip output. The strict separation of test and live data via the quality bits ensures that maintenance activities cannot inadvertently trigger a real breaker trip.