Commercial Solar Plant Performance Monitoring in Bangladesh: PR, Availability & O&M KPI Guide
Direct answer: A commercial solar plant in Bangladesh is healthy when three KPIs stay in range — Performance Ratio (PR) of 75–85% for rooftop systems measured per IEC 61724-1, energy-based availability of 98–99%, and specific yield of roughly 1,200–1,400 kWh per installed kWp per year. Track all three monthly through your inverter portal plus an onsite pyranometer. A sustained PR drop of 5 percentage points or more means soiling, shading, a dead string, or an inverter fault — investigate it, do not accept it as normal variation.
Why monitoring is an ROI tool, not a luxury
A 200 kWp factory rooftop plant in Gazipur at Bangladesh's typical 3.8–4.2 peak sun hours should produce roughly 650–700 kWh per day. If no one watches the data, a failed string fuse or a drifted MPPT setting can quietly erase 10–20% of that output for months. At industrial tariffs, even a 5% undetected loss on a 200 kWp plant means roughly 40,000–50,000 BDT of electricity value lost every month — far more than the cost of a data logger and a quarterly inspection visit.
KPI 1: Performance Ratio (PR) — the plant's health score
Definition (quotable): Performance Ratio is the actual AC energy a solar plant delivers, divided by the energy it would deliver if it converted every photon of measured sunlight at nameplate efficiency — expressed as a percentage after normalising for the plant's installed capacity (IEC 61724-1).
In plain terms: PR tells you what fraction of the available sunshine actually became billable electricity, after accounting for heat, dust, cable losses, inverter losses, and downtime.
Healthy bands for commercial rooftop systems:
| PR range | Meaning |
|---|---|
| Above 80% | Efficient, well-maintained plant — losses inside expected limits |
| 70–80% | Normal for most commercial plants once temperature, moderate soiling, and inverter losses are included |
| Below 70% | Investigate — persistent shading, heavy soiling, dead strings, or inverter problems |
Worked example (Dhaka, 100 kWp): On a day with 4.0 peak sun hours, the theoretical yield is 100 × 4.0 = 400 kWh. If the plant delivers 320 kWh, PR = 320 ÷ 400 = 80% — a healthy result for a Bangladeshi rooftop in summer. Compare PR month-to-month, not day-to-day, and always compare against the plant's own history, not a national sticker target.
KPI 2: Availability — is the plant actually running?
Definition: Energy-based availability is the fraction of theoretically producible energy the plant actually delivered, excluding periods the grid was unavailable (curtailment, load-shedding outages).
Modern O&M contracts guarantee a minimum of 98% energy-based availability per year (Greensolver industry best practice), rising to 99% in many utility contracts. Time-based availability ("the inverter was on") is a weaker metric — a plant can be "on" while a string is dead. Insist on energy-based availability in any O&M agreement; it is the metric that tracks money, not hours.
KPI 3: Specific yield — kWh per kWp per year
Specific yield normalises annual production by installed capacity, letting you compare a 50 kWp and a 500 kWp plant fairly. For Bangladesh, expect roughly 1,200–1,400 kWh/kWp/year for a well-oriented, unshaded rooftop plant — higher on clean, cool winter days, lower in dusty April or monsoon cloud cover. If a commissioned plant delivers far below this range after its first year, demand a loss analysis before the EPC warranty period expires.
What a real monitoring stack looks like
- Inverter cloud portal (SolisCloud, Huawei FusionSolar, Sungrow iSolarCloud): string-level current/voltage graphs — your first screen every month.
- Data logger / gateway with 4G connectivity: stores 5–15 minute interval data even when the internet drops.
- Onsite pyranometer (plane-of-array): without measured irradiance, PR is a guess. IEC 61724 requires onsite measurement for any PR you intend to argue about.
- Module temperature sensor: lets you compute temperature-corrected PR, which stays flat across seasons and isolates real faults from summer heat.
- For plants above ~500 kW: a SCADA dashboard aggregating all inverters, the weather station, and the export meter in one place.
Keep the pyranometer clean and recalibrated (manufacturers typically recommend a 2-year cycle) — a dirty reference sensor makes a healthy plant look sick.
Reading alarms like an O&M professional
| Symptom | Likely cause | First action |
|---|---|---|
| One string current far below its siblings | Blown fuse, MC4 connector corrosion, PID | Isolate and inspect the string; thermal scan the connectors |
| Midday clipping on a flat plateau | Inverter undersized vs array (high DC/AC ratio) or grid export limit | Check whether clipping was designed-in (see our hybrid inverter sizing guide) |
| All strings drift down together over weeks | Soiling (dust) or seasonal haze | Compare with the soiling losses in our rooftop maintenance guide — clean a reference block and re-measure |
| PR fine in winter, collapsing every summer | Temperature losses — normal physics, not a fault | Switch reporting to temperature-corrected PR |
| Zero output, grid voltage present | Inverter fault, AC breaker trip, or isolation | Check fault codes; verify DC isolators and AC protection |
| Zero output during load-shedding | Anti-islanding shutdown (<2 s) — mandatory per SREDA rules | Not a fault. Full detail in our net metering guidelines |
A healthy plant's PR falls gradually (ageing is typically under 0.5% per year for modern modules) — a sudden drop is always an event, never ageing.
What to demand in an O&M contract
Greensolver's industry survey of O&M contracts flags a recent shift: asset owners are moving from PR guarantees toward Response Time (RT) guarantees — the time the contractor takes to acknowledge an alarm, investigate, and fix it, tiered by fault severity. For a Bangladeshi factory plant, a strong contract should include:
- Energy-based availability ≥ 98% per operational year, with liquidated damages below it.
- Response time tiers: e.g. critical faults (full plant down) acknowledged within 4 hours; string-level faults within 48 hours.
- Scheduled preventive visits: quarterly inspection per IEC 62446-1, plus cleaning frequency matched to your site's dust load.
- Reporting: monthly PR, availability, and yield report with irradiance data attached — not just a screenshot.
- Spare parts: critical spares (fuses, connectors, one spare inverter or fast replacement SLA) on the contractor's account.
FAQs
How often should I actually check my plant's PR?
Monthly is the professional standard — modern portals calculate it automatically and can email you when PR deviates from the expected band. Weekly glances at string graphs catch dead strings faster.
What PR is "good" for a rooftop plant in Bangladesh?
Above 80% is excellent and typical of clean, well-ventilated plants; 70–80% is normal once heat and moderate dust are included. Sustained PR below 70% deserves a structured loss investigation, not patience.
Can PR be compared between summer and monsoon months?
Only with temperature-corrected PR. Conventional PR naturally dips in hot months because module voltage drops with temperature — that is physics, not a fault. Temperature correction removes the seasonal swing so real problems stand out.
Do I really need a pyranometer, or can I use weather data?
Satellite or nearby-station irradiance is fine for planning, but PR calculated without onsite irradiance is unreliable — a 5% irradiance error becomes a 5% PR error. For any performance guarantee or dispute, IEC 61724 expects an onsite measurement.
Sources: IEC 61724-1 (PV system performance monitoring); Greensolver — Key O&M contract elements (98% energy-based availability, response-time guarantees); maysunsolar / kindastuff PR benchmark bands for rooftop systems (75–85%, >80% efficient); NREL — module degradation typically under 0.5%/year.
Need Commercial Equipment for Your Project?
Get manufacturer-direct wholesale pricing, complete compliance documentation, and Dhaka warehouse delivery lead times.