Your app tells you the plant is working. We tell you which string isn’t.
Most solar monitoring shows you one number for the whole plant. If one string in twenty fails, that number barely moves — so nothing looks wrong, and you lose that output every sunny day for months. A single dead string is invisible on your inverter app, invisible on your bill, and invisible on the plant total.
We watch every string individually. When one stops, we know that day — not at the end of the quarter when the numbers finally don’t add up. In the last 60 days, 61% of the inverters we monitor had a dead string or stopped reporting entirely — on systems whose owners were certain everything was fine.
The numbers behind it
None of these is a projection or a model. Each is measured on the live platform, over a stated period, and stated conservatively.
- The 31% is understated if anything — the “after” window falls later in the season, when sunlight is weaker.
- The 2,610 MWh is measured on our platform since April 2026 — roughly 1,000 tonnes of CO₂ avoided at Pakistan’s grid carbon intensity.
| Finding | Measured |
|---|---|
| Inverters with a dead string or that went silent (60 days) | 61% |
| Output recovered when one dead string on a four-string inverter was fixed | +31% (8 days after repair) |
| Reduction in alert noise after our de-duplication fix | 76% — the same faults, a fraction of the messages |
| Clean solar generation measured on our platform since April 2026 | 2,610 MWh |
| Uptime across the systems we monitor (60 days) | 98% |
The scale we watch at
This is not a pilot. BijliBachao’s Solar Performance Cloud continuously inspects solar across Pakistan, string by string, and the numbers grow every month.
- Every major inverter brand in Pakistan — Huawei, Solis, Growatt, Sungrow, Canadian Solar, Solarman, GoodWe — and counting.
- From 2 MW installations down to single rooftops.
- String-level, not plant-level — the category claim nobody else in this market makes.
| Measure | Value |
|---|---|
| Solar strings under continuous inspection | Over 1,000 |
| Sites across Pakistan | 90+ |
| Inverters monitored | 100+ (seven manufacturers) |
| Measurements taken | More than 100,000 every day |
| Uptime across monitored systems | 98% — the other 2% seen the day it happens |
| Fault to verified alert | Minutes, not months |
Five real things we found
Every one of these is a genuine incident from the systems we monitor, with names and locations removed. Three of them are not even faults — they are things nobody was looking for, which is exactly why independent inspection found them.
“The plant looked fine. Five inverters were dead.” At a mill we monitor, five inverters stopped reporting entirely. The manufacturer’s own cloud platform raised nothing at all. Our silence alert fired the same day.
“They already owned the expansion.” At one industrial site, the inverters carry dozens of unused string inputs — enough for roughly double the panels currently installed. The equipment was fitted years ago and has sat idle ever since. Nobody had looked.
“One tracker at a third of the voltage of its twin.” Two strings on the same inverter: one at 493 volts, the other at 175. Missing panels or a broken connection. The system had never been monitored at string level, so it had never shown up.
“The inverter reported its own fault. Nobody was listening.” A commercial rooftop’s inverter flagged an internal PID fault twice in one afternoon. The fault message existed the whole time — it just had nowhere to go until it reached our platform.
“The paperwork was half the truth.” One site’s asset register understated its own solar system by more than half — found by comparing the records against what the panels actually produce.
What we catch — and what we read
We find what your inverter never tells you — and we also read every fault it does report, in plain language, across seven brands. Those are two distinct capabilities, and both matter.
- We detect these ourselves, from our own analysis of string-level data: dead strings, silent inverters, and strings underperforming against their peers on the same inverter.
- We surface these from the inverter’s own fault reporting, in plain language: insulation faults, fan alarms, phase faults, grid loss, PID faults, residual-current (earth leakage), grid over/under-voltage and frequency, and string reverse connection.
Why plant-level monitoring misses all of this
You can’t manage what you can’t see. An inverter app reports its own hardware’s numbers, one brand at a time, and mostly tells you what already happened. It shows “green lights” — the system is on — not lost revenue. The failure mode that costs the most is precisely the one it misses: a single string quietly underperforming while the plant total still looks normal.
Independent, string-level, multi-brand inspection is the opposite of that. It is not tied to your inverter brand, it watches each string on its own, and it turns what it finds into a prioritised action — AI flags and ranks the issue, and solar engineers verify it before you hear about it. The AI recommends; people decide.
