In centralized protection and control (CPC), the protection and control functions of several bays, or of the whole substation, run on one or two high-performance computing platforms instead of one IED per bay. Measurements arrive as sampled values from merging units over the process bus, and trips go out over GOOSE. Availability depends on duplicating the platform, the network and the time source.
What centralized protection and control is
In a conventional substation each bay has its own protective relay and its own bay controller. Each IED receives the currents and voltages of its instrument transformers, computes its functions and trips its breaker. The intelligence is distributed.
Centralized protection and control (CPC) changes that split. The protection, control and monitoring functions of several bays, or of the whole substation, run on one or two high-performance computing platforms. Measurements arrive digitized from the switchyard, and commands go back over the same network.
Working Group K15 of the IEEE PES Power System Relaying and Control Committee (PSRC) published the report Centralized Substation Protection and Control in December 2015. It describes the state of the art and compares the traditional scheme with the centralized one in asset management, maintenance, security, interoperability and the station interface. It is the starting reference for any specification.
Where the measurements come from: the process bus
CPC is only possible in an IEC 61850 digital substation. In the switchyard, a merging unit takes the signals from the current and voltage transformers and publishes them as sampled values (SV), according to IEC 61850-9-2 and the digital interface of IEC 61869-9. Breaker and disconnector positions travel as GOOSE, and trips return to the switchyard the same way, to the device that drives the trip coil.
That network is the process bus. In a distributed scheme each IED subscribes to the streams of its own bay; in a centralized one, the platform subscribes to the streams of every bay. Bandwidth, priority and segmentation of the bus become first-order design decisions.
- 01Merging unitDigitizes the bay current and voltage and publishes sampled values
- 02Process busCarries the streams of every bay, duplicated with zero-recovery-time redundancy
- 03CPC platformComputes the protection functions of every bay with synchronized samples
- 04GOOSE tripReturns to the switchyard over the process bus, to the faulted bay
- 05Circuit breakerOpens on the command of the switchyard device that drives its coil
Claimed advantages
- Fewer devices to maintain. One platform instead of many IEDs reduces the number of different firmware versions, configurations and spare parts.
- Data from the whole substation in one place. Functions that need information from several bays, such as 87B bus differential or breaker failure, are solved without exchanging messages between devices.
- Changes by software. Adding a bay or a function is a configuration task on the platform, provided it has spare computing capacity.
- Separate life cycles. The switchyard (merging units, short wiring) has a long life; the computing platform can be renewed without touching the switchyard.
None of these advantages is automatic. They depend on how redundancy, cybersecurity and testing are solved.
Redundancy: the point that defines the design
In a distributed scheme, the failure of one relay leaves one bay without main protection, and that bay keeps its backup. In a centralized scheme, a platform failure affects every bay at once. That is why the design starts by removing single points of failure:
- Two independent CPC platforms, each with every function, or one main platform with a different backup protection, following the same logic as main 1 and main 2 protection.
- Duplicated process bus with PRP or HSR (IEC 62439-3), so that losing a link or a switch does not interrupt the samples.
- Redundant time synchronization with PTP according to IEC/IEEE 61850-9-3. If the samples of two merging units lose their common reference, the functions that compare them must block.
- Separate power supplies and DC station service for each platform.
Some designs also keep a minimal local protection in the switchyard or in the bay, so that a common-mode failure of the central part does not leave the substation unprotected.
Technology status
The distributed scheme is still general practice in high and extra-high voltage. Centralized protection, and its virtualized variant, where the functions run as software on generic servers, are in early adoption. Before choosing it, ask the manufacturer for references of installations in service, the documented behavior on loss of samples or synchronization, and the software update policy over the life of the substation.
The relationship with the manufacturer changes as well. With one IED per bay, each relay comes with its IEC 61850 conformance certificate and its service history. With a central platform, reliability depends on the hardware, the operating system, the virtualization layer if there is one, and the protection application, and each layer may have a different owner.
What changes in testing
- No conventional injection into the platform. There are no analog inputs: the test is done by publishing simulated sampled values on the process bus, with the device in IEC 61850 test or simulation mode.
- Network tests as part of the protection test: loss of a link, loss of a merging unit, loss of synchronization and bus load.
- Isolating one bay without affecting the others, because they all share the platform.
- Regression after every software change. A platform update touches every bay, so the function tests have to be repeated systematically.
When samples are lost in service, the diagnosis follows the same path as in any process bus; we explain it in sampled values lost on the process bus.
Distributed or centralized: how to decide
- Does the substation have, or will it have, a process bus with merging units?
- NoThe distributed scheme with one IED per bay remains the natural choice
- YesCan the project duplicate the platform, the process bus and time synchronization?
- YesCPC can be evaluated with an availability analysis and manufacturer references
- NoKeep one IED per bay on the process bus
We help you evaluate it
If your project includes a digital substation and you want to compare a distributed scheme with a centralized one, Tiza Group does the secondary engineering: protection architecture, process bus redundancy, synchronization and test plan, for high and extra-high voltage substations and for industry. If the IEDs are already on site, we configure them through remote relay configuration. Contact us.
Related guides
- IEC 61850 digital substation: process bus, GOOSE and SV
- IEC 61850 process bus: SV, GOOSE, timing and bandwidth
- Merging unit: IEC 61869-9, 9-2LE and testing
- PRP and HSR (IEC 62439-3): zero-recovery-time redundancy
- Main and backup protection
- Bus differential 87B (PDIF): what it is and how it works
- Line protection main 1 and main 2: how it is built
- Sampled values lost on the process bus
- PTP IEEE 1588 and the IEC/IEEE 61850-9-3 power profile
- Sampled Values and GOOSE bandwidth: how to estimate it
- Substation network redundancy: RSTP, PRP, HSR or MRP
- PTP, NTP or IRIG-B: which synchronization each application uses
- Substation GNSS clock: holdover, redundancy and risks
References
- IEEE PES Power System Relaying and Control Committee, Working Group K15, Centralized Substation Protection and Control, report, December 2015.
- IEC 61850-9-2: Sampled values over ISO/IEC 8802-3.
- IEC 61869-9: Digital interface for instrument transformers.
- IEC 62439-3: Parallel Redundancy Protocol (PRP) and High-availability Seamless Redundancy (HSR).
- IEC/IEEE 61850-9-3: Precision time protocol profile for power utility automation.
Frequently asked questions
Does centralized protection remove the relays from each bay?
It can, but many designs keep a local backup protection in the critical bays, because a common-mode failure of the central platform would leave the whole substation unprotected. The decision comes from the availability analysis of the project.
What is the difference between centralized and virtualized protection?
Centralized protection gathers the functions of several bays on one platform. Virtualized protection is one way of doing it, where the functions run as software on a generic server with a virtualization layer instead of on dedicated manufacturer hardware.
What does an existing substation need to move to centralized protection?
It needs to digitize measurements and commands in the switchyard with merging units, a process bus with zero-recovery-time redundancy and PTP time synchronization. Without that base, the central platform has nowhere to take its samples from.