A tripped string
One string stops. The plant still produces, so nobody looks.
03 Catch
Catch losses early.
Know the moment a plant makes less than it should, and who is on it.
Why it matters
The costly faults are the quiet ones.
A plant that stops gets noticed, but a plant that makes a little less can go unseen for weeks.
A dirty array
Output slides a little every day. No single day looks wrong.
A restarting inverter
It drops out every afternoon and comes back. The daily total hides it.
GridMarshal compares each part of the plant with what it should make, so small losses show up while they are still small.
Rules and escalation
Set the rule once. Use it everywhere.
Alarm templates carry your rules across the fleet, with escalation built in.
01
Template
Define an alarm once, with the condition that opens it.
02
Rule
Apply it to the plants and equipment it should watch.
03
Alert
The right person hears about it on the channel they use.
04
Escalation
If it needs more attention, it moves up the chain you set.
Expected output
What it made, against what it should.
Each solar inverter has an expected output, so you know when actual falls away from it.
- Expected output worked out per inverter
- Actual output compared with it, live
- Deviation alarm when the gap opens
- Forecast next to production in monitoring
String comparison
The neighbour test.
Strings on the same inverter see the same sun, so they should make about the same current.
- Each string compared with the mean of its neighbours
- Within 5 percent is in line
- Up to 10 percent is worth a look
- Above 10 percent is losing energy now
Tripped fusesSoilingShadingDamaged modulesLoose connectors
Common causes a low string can point to.
Communication alarms
Silence is an alarm too.
When a plant stops sending data, you hear about it.
Data stops
A logger goes quiet. A communication alarm opens.
You know which kind
A data problem is kept apart from a production problem, so the right person goes.
History comes back
When the logger reconnects, missing history can be back-filled.
Alert channels
Three ways to hear it.
Each person gets alerts for the plants they are given, where they will see them.
A clear record in the inbox, with the plant and the alarm named.
Telegram
Straight to the chat app your field team already uses.
Phone push
From the GridMarshal phone app, on Android and iPhone.
Calm under alarms
One alert. A clear owner.
The right person gets the right alert once, with what to do next.
- Open alarms on every plant in the portfolio view
- Turn an alarm into a ticket with a name on it
- Energy lost recorded on the ticket
Questions
Alarms, answered.
What teams ask before they switch their alerts over.
Who receives an alarm?
Only people who are given that plant. Rules and escalation decide who hears about it first and who hears next.
What happens if a data logger stops sending?
A communication alarm opens. When the logger reconnects, missing history can be back-filled.
Which alert channels are supported?
E-mail, Telegram and push notifications from the phone app.
Can an alarm become a work order?
Yes. An alarm can become a ticket with an owner, a workflow and the energy lost. Operations and maintenance
Catch losses early
See alarms on your own plants.
A walkthrough of alarms and expected output, with your own plants in mind.