Control rooms
What Belongs in a Substation Control Room, and What Does Not
By Arch. Dany Dandachi, ALMAFOR, Abu Dhabi, United Arab Emirates
Published
A control room is not judged on the day it is commissioned. It is judged at three in the morning, by somebody who did not build it, during an event nobody planned for.
Almost everything worth deciding about one follows from that.
What is actually in the room
Strip the furniture and a station control room is six things.
The supervisory servers. Usually a pair, because the station's picture of itself should not depend on one machine. They collect from the bay controllers, hold the historical record and serve the displays.
The operator position. One or two screens showing the single line diagram, the alarm list and the event log. This is the surface the station is actually run from, and it should be difficult to misuse.
The engineering position, separately. Configuration, logic, addressing and diagnostics. It belongs on its own machine, with its own access, because the alternative is that every configuration change gets made on the workstation somebody is using to operate a live network.
The gateways. What publishes the station outward. On the Musaffah station that meant redundant pairs, one publishing the 220 kV side to the load dispatch centre and another publishing the 33 and 11 kV side to the distribution management system: two audiences, two data sets, deliberately not served from one device.
The clock. A GPS reference that disciplines every device in the station. Without it the sequence of events is a record of when packets arrived, not of what happened first, and a disturbance investigation cannot reach a conclusion.
The UPS. Sized for the outage the station has actually had. The control room is the instrument that reports the power failure; it has to survive the failure to do so. Our TRANSCO stations carry 5 and 6 kVA units for exactly this.
The part most people get wrong
The alarm list, and it is not close.
A control system can raise an alarm on every point it reads. Do that and the screen is permanently red, the operator correctly learns that red means nothing, and the one alarm that mattered arrives into a room that has stopped looking.
The list should be built backwards from the decisions an operator has to make. What do they need to know in order to act, what can wait for the morning report, and what is diagnostic data that belongs in the historian and nowhere else. That is an operations conversation, not an engineering one, and it is worth having before the graphics are drawn.
The same discipline applies to the mimic. If the single line on the screen does not match what a person sees when they walk the switchgear, they will stop trusting it, and after that the room is furniture.
Keep the hardwired layer
There is a temptation, on a modernisation, to sweep away the hardwired mimic and interlocking because the new system can do all of it in software.
At Al Khazna we upgraded the upper and lower ring control and monitoring system at 33 and 11 kV and deliberately retained the hardwired interlocking through the change. The reason is simple: during the cutover, the new system is exactly the thing that cannot be trusted yet. A station whose safety depends on the equipment being installed has no safe way to install it.
Wired interlocking also keeps working when the supervisory layer is down, being patched, or being replaced in ten years by somebody who has not been born into this project. It is slower and less elegant, and it is why the work can be done on a live station at all.
The room is part of the contract
Control rooms are frequently treated as somebody else's scope, which is how a station ends up with a well engineered system in a room with no cable route to it.
At Khalifa Port the scope across three primary substations included the site survey, detailed design, shop drawings, civil works, and the IT and fibre network alongside the hardware and software replacement. On TRANSCO project N-15223 the same pattern held across six grid stations from 400 kV down to 11 kV, with the control room equipment, the UPS, the fibre and the control centre integration all in one contract.
When the room, the network and the system are split between contractors, every intermittent fault becomes a meeting.
A short specification checklist
- Operator and engineering positions on separate machines with separate access
- Redundant servers; single points of failure identified and accepted deliberately, not by accident
- GPS time synchronisation, with timestamps taken at the bay and not at the server
- UPS autonomy set against the longest outage the station has actually seen
- Alarm list derived from operator decisions, then reviewed after six months in service
- Mimic that matches the plant as built, not the plant as drawn
- Hardwired interlocking retained where the station cannot be taken out of service
- As built drawings, configuration backups and a named owner at handover
Working with us
Almafor designs, builds and commissions substation control and monitoring systems and the rooms they run from, on live transmission and distribution stations from 400 kV down to 11 kV, including the civil and fibre works. Clients include TRANSCO, TAQA Distribution, Al Ain Distribution Company and AD Ports Group.
Call +971 55 954 5801 or email info@almafor.com.
Common questions
Questions we are asked about this
What equipment goes in a substation control room?
A supervisory server pair, an operator workstation, a separate engineering workstation, the gateways that publish the station to a control centre, a station clock for time synchronisation, network switches and fibre patch panels, and the UPS that keeps all of it alive when the auxiliary supply fails. Older stations add a hardwired mimic panel, which is often worth keeping.
Why separate the operator and engineering workstations?
They do different jobs and carry different risk. The operator position runs the station and should be hard to misuse. The engineering position changes configuration, logic and addressing, and should be hard to reach. Combining them means every configuration change happens on the machine somebody is using to run a live network.
Is a hardwired mimic panel obsolete?
No, and we have deliberately retained hardwired interlocking through an SCMS upgrade. A mimic that is wired rather than rendered keeps working when the supervisory layer is down, being patched, or being replaced. On a station that cannot be taken out of service it is the thing that lets the work happen at all.
How many alarms should a control room have?
Far fewer than the equipment can generate. The right number comes from the decisions an operator actually has to make, not from the I/O schedule. A list built from the schedule produces a screen that is permanently red, and an operator who has correctly learned to ignore it.
Does the control room need its own UPS?
Yes, and sized for the outage the station has actually experienced rather than the one in the specification. The control room is the instrument that tells you the power failed, so it has to still be running to say so. Stations we have delivered carry 5 and 6 kVA units for this.
We want to hear from you
Tell us about the system you need built, upgraded or kept running.