Most factory owners we meet worry about whether solar will work. The better question, a few years in, is whether it's still working as well as it should. A plant that was designed to run at, say, an 80% performance ratio can quietly settle into the low 70s without anyone raising an alarm, because nothing actually breaks. The panels are still there. The inverter still hums. The bill is still lower than it used to be. It just isn't as low as it ought to be, and that gap is money leaking out of an asset you've already paid for in full.
This is a plain-English guide to why that happens and what to do about it, written for plant heads and finance folks in Tamil Nadu and South India. Every figure here is indicative and reviewed September 2026: real numbers move with your site, your modules and the season. But the physics doesn't change, and neither does the arithmetic: in a state where daytime power on an HT connection costs roughly ₹8–11 a unit, every unit your plant fails to make is a unit you buy back from the DISCOM at retail. If you're still weighing the economics of going solar in the first place, our companion piece on rooftop solar for factories in Tamil Nadu covers cost, savings and payback; this one is about protecting that return once the plant is on the roof.
What performance ratio actually measures
Performance ratio (PR) is the energy your plant actually delivered divided by what it should have delivered given the sunlight it received: expressed as a percentage, it's the single most honest scorecard of plant health. Because it's normalised against the irradiance the plant genuinely saw, a cloudy week and a blazing one are judged fairly: PR asks "of the sunshine you got, how much did you convert?" not "how sunny was it?"
A well-designed, well-kept C&I rooftop plant in South India usually runs a PR of about 78–82% indicative. Left alone (no cleaning, faults ignored) the same plant can drift to 68–72%, and the owner rarely notices because the decline is gradual and the bills are still lower than the pre-solar days. That 10-point slide is the whole subject of this article.
PR is easy to confuse with CUF, the capacity utilisation factor, so it's worth separating them. CUF is your annual generation divided by what the plant would make running flat-out every hour of the year, and in Tamil Nadu it typically lands around 16–19% indicative, not because the plant is unhealthy, but because the sun only shines part of the day. CUF measures the resource; PR measures the plant. A beautifully maintained plant will still show a modest CUF, so never let a low-sounding CUF worry you. PR is the number that tells you whether the plant itself is doing its job.
| Performance ratio | What it usually means | Typical for |
|---|---|---|
| 80–85% | Excellent: new, well-designed and well-kept | New plants on strong sites |
| 78–82% | Healthy: normal for a maintained plant | The target for most rooftops |
| 73–77% | Slipping: soiling or minor faults building up | A plant overdue for attention |
| 68–72% | Neglected: real money lost every day | Plants running without O&M |
Where a plant loses generation
Between the nameplate rating printed on your modules and the units that reach your meter sits a stack of small deductions, and no single one of them looks alarming on its own. That's exactly why they're dangerous: each is a percent or two, easy to shrug off, but they stack up and compound across a year until a fifth of your potential generation has quietly gone missing.
The list below is the usual cast. Notice that some are one-time, fixed facts of physics (conversion, mismatch), some creep in with age (degradation), and some are entirely within your control (soiling, downtime). It's the controllable ones that separate an 80% plant from a 72% one.
| Loss source | Typical range | What controls it |
|---|---|---|
| Soiling (dust, pollen, droppings) | 3–6% (far higher if neglected) | Cleaning cadence & water quality |
| Heat / temperature derating | 5–9% | Module tech, ventilation, mounting height |
| Wiring & inverter conversion | 3–5% | Cable sizing, inverter quality & sizing |
| Downtime & tripping | 1–4% | Monitoring & fast fault response |
| Module mismatch & tolerance | 1–2% | Binning & sensible string design |
| Annual degradation | 0.4–0.7% per year | Module quality & PID control |
These ranges are indicative and they interact (losses multiply rather than simply add) but the shape is always the same. Design and component quality set the floor; O&M decides whether you stay near it or drift away from it. Two of these losses deserve a closer look, because they're the ones a plant in our climate feels hardest: soiling and heat.
Soiling: the South India tax
In South India, dust isn't a nuisance: it's a tax on your generation, and it's the single loss you have the most control over. A film of dust, pollen, cement or quarry particulate, salt near the coast, and the inevitable bird droppings all block light before it ever reaches the cells. What makes soiling insidious is that it's invisible from the ground and completely gradual: the plant looks fine, and it's steadily making less.
In the dry pre-monsoon months, an uncleaned rooftop plant here can shed roughly 1–1.5% of output per week indicative, and a plant that's genuinely neglected through a full year, especially near industrial dust, can give up 15–25% of its annual generation indicative to soiling alone. That's not a rounding error; that's a quarter of your solar investment sitting under a layer of dust.
The fix is unglamorous but decisive: clean on a sensible cadence. For most Tamil Nadu rooftops that means every two to four weeks, tightening during the dusty, rain-free stretch from roughly January to June, and tighter still for plants downwind of cement works, quarries or an unpaved road. A monitored plant can do better still: clean when the data shows soiling has crossed the point where a cleaning visit pays for itself, rather than blindly by the calendar.
How you clean matters as much as how often. Use low-TDS or DM (demineralised) water where you can: hard borewell water leaves its own mineral haze once it dries, so you can "clean" a panel and leave it dirtier. No abrasive pads, no harsh detergents, and never a pressure jet aimed at the module edges or junction boxes. And none of it is worth an injury: rooftop cleaning needs proper anchors, edge protection and trained hands, not a casual worker with a bucket. Where water is scarce or roofs are hard to reach, robotic and waterless (microfibre) cleaning systems are increasingly practical, though they suit larger or ground-mount plants more than small rooftops.
Soiling is the only major loss you can reverse with a bucket of clean water, which is precisely why leaving it uncleaned is the most expensive habit a plant can have.
Heat, degradation and long-term drift
Solar modules are rated at 25 °C in a lab, but a Tamil Nadu rooftop almost never sees 25 °C, so real output always sits a little below the sticker on the back. Panels are counter-intuitive that way: they love light but dislike heat, and on a hot afternoon a module's cells can easily run 55–65 °C even as they make good power.
The number that governs this is the temperature coefficient, typically around −0.34 to −0.4% per °C indicative for modern mono-PERC modules. Read it plainly: for every degree the cell sits above 25 °C, you lose roughly a third of a percent of output. Over the 30-degree gap between the lab rating and a real rooftop cell, that's a meaningful, permanent daytime derating, and it's why the same module can genuinely out-generate on a cool, bright morning in December than on a scorching May afternoon.
You can't change the weather, but design blunts the heat. Modules mounted flush against a hot metal roof cook; a little mounting height and a clear air gap let them breathe and shed heat, which is one quiet reason a thoughtfully engineered plant out-generates a cheap one on identical panels. On top of the daily heat effect, modules degrade slowly with age: expect a first-year settling of about 1–2% (light-induced degradation), then a long, gentle slope of roughly 0.4–0.7% per year indicative for good mono-PERC. That drift is normal and predictable: the danger is confusing it with soiling or a fault, which is where honest monitoring earns its place.
Not sure whether your plant is drifting or just dirty?
Ask for a performance reviewWhat good O&M actually does
Good O&M isn't a watchman with a broom: it's the discipline that stops every one of those small losses from quietly becoming permanent. The best operators work to protect one number, your PR, and by extension the rupees behind it. Here's what that discipline looks like in practice, and why each piece maps directly back to generation you'd otherwise lose:
- Remote monitoring and alerts: the foundation. If a string drops out at 9 a.m., you want to know by 9:05, not on next month's bill. Fast visibility is what turns a week of lost units into an hour of them.
- Scheduled and condition-based cleaning: a fixed cadence, tightened by what the data actually shows, so you clean when it pays and not a rupee sooner.
- String and inverter diagnostics: comparing strings against each other to catch the underperformer before it becomes a dead zone, and watching inverters for nuisance trips and clipping.
- IV-curve tracing and thermography: an IV trace fingerprints a struggling string, and an infrared scan reveals hotspots, cracked cells and loose connections you'd never see by eye.
- PID and connector checks: testing for potential-induced degradation, and inspecting DC connectors for the loose, arcing joints that are both a generation loss and a genuine fire risk.
- Vegetation and shade-creep control: the tree that wasn't a problem at commissioning is a problem three summers later; a shaded string drags down more than its own share.
Underneath all of it sits a simple hierarchy. Corrective maintenance fixes what's already broken. Preventive maintenance is the routine (cleaning, inspections, tightening) that stops breakage in the first place. Predictive maintenance uses the monitoring data to act on a trend before it becomes a fault at all. Cheap operators do only the first; good ones live in the second and third, because that's where PR (and your return) is actually defended.
Key takeaways
- PR is the scorecard. A maintained C&I rooftop should hold ~78–82%; a slide into the low 70s is money leaking, not bad luck.
- Soiling is the controllable loss. Clean every 2–4 weeks, tighter in the dry season, with low-TDS water and no abrasives.
- Heat and degradation are physics. You can't remove them, but ventilation, mounting and quality modules keep them small and predictable.
- Monitoring turns weeks into hours. The fastest ROI in O&M is simply knowing a fault happened the day it happened.
- O&M protects an asset you already own. It's not an extra cost; it's what keeps the one you paid for performing.
Reading an O&M contract
A bankable O&M contract puts numbers on paper (a PR or generation guarantee, response times, a cleaning schedule) so your plant's performance isn't left to goodwill. When you're comparing offers, the price per year tells you almost nothing on its own; what it buys is everything. A good agreement should spell out most of the following:
- A PR or generation guarantee: a committed performance level, with a mechanism if the plant falls short. This is what separates a service from a promise.
- Response-time SLAs: how fast a fault gets acknowledged and how fast someone is on the roof. A one-day response and a two-week response are different products entirely.
- A defined cleaning frequency: stated in the contract, with seasonal tightening, not left to "as required".
- Spares and inverter support: who holds critical spares, and what happens the day an inverter fails out of warranty.
- A reporting cadence: monthly generation and PR reports you can actually read, so the plant's health is never a mystery.
Be wary of "cheap O&M" that turns out to be one annual visit and a phone number. It looks like a saving until the season a string sits dead for two months and eats far more in lost units than a proper contract would have cost for the year. The same discipline you applied when you weighed what the plant cost to build belongs here too: look at what's included, not just the number. If you want to see how maintained plants actually behave over time, our case studies walk through real sites across South India.
And if you're still at the sizing stage, remember that O&M economics start at the design table: a plant sized honestly to your daytime load, ventilated properly and wired sensibly is simply cheaper to keep healthy. Our quick solar calculator is a good first-cut on what your roof and load could support.
Frequently asked questions
How often should solar panels be cleaned in Tamil Nadu?
For most rooftop plants in Tamil Nadu, every two to four weeks is a sensible baseline, and you should tighten that to weekly or fortnightly during the dry, dusty months from roughly January to June, or if you're near cement works, quarries or a busy road. The right cadence is the one where the cost of a cleaning visit stays below the value of the generation you'd otherwise lose, so a monitored plant can clean on evidence rather than a fixed calendar.
What is a good performance ratio for a rooftop plant?
A healthy commercial rooftop plant in South India typically runs a performance ratio of about 78 to 82 percent. If yours has drifted into the high 60s or low 70s, that's a clear signal something is wrong (usually soiling, an underperforming string, or an inverter that keeps tripping) and it's costing you real units every day it stays uncorrected.
Does O&M come included when I buy a plant?
Usually only for the first year. A good turnkey EPC contract includes first-year operations and maintenance, but after that O&M is a separate ongoing service you either renew with your installer or hand to a specialist. Treat it as essential, not optional: an unmaintained plant quietly loses generation every year it runs without proper care.
How much generation can poor O&M cost me over 25 years?
Over a 25-year life it's easily the difference between a plant that averages around 80 percent PR and one that drifts to the low 70s, a gap of roughly 10 percentage points of lifetime output. On a plant that would otherwise offset power worth ₹8 to ₹11 a unit, that lost generation runs into tens of lakhs of rupees, which is why O&M is best seen as protecting an asset you've already paid for rather than an extra cost.