Solar irradiance and degradation modelling for PV systems in East Africa coastal climates
Energy yield models calibrated on temperate-climate defaults systematically overstate the long-term output of PV systems on the East African coast. This brief explains which assumptions drift, by how much, and how to correct a PVsyst model so that financial projections survive contact with the site.
Where the defaults go wrong
Three default assumptions dominate the error. Module degradation defaults of 0.5 percent per year reflect temperate field studies; in high-humidity coastal environments at sustained high temperature, observed degradation for standard modules runs 0.8 to 1.2 percent per year, driven by accelerated encapsulant browning and potential-induced degradation. Soiling defaults of 2 to 3 percent understate coastal conditions where salt spray combines with dust; without an aggressive cleaning regime, sustained soiling losses of 5 to 8 percent are common in our metered data. And thermal loss parameters calibrated for free-standing arrays in moderate ambient underestimate cell temperatures reached in 38 to 42 degree coastal conditions, particularly for rooftop mounting with restricted airflow.
Calibrating the model honestly
Our practice is to model two scenarios and finance against the conservative one. The base case uses regionally evidenced parameters: degradation at 0.9 percent per year for standard glass-backsheet modules, or the manufacturer's warranted curve where glass-glass modules with PID-resistant cells are specified; soiling at 6 percent with a defined cleaning schedule, or 3 percent where cleaning is contracted and budgeted explicitly; and thermal loss coefficients adjusted to measured local ambient and the actual mounting configuration.
Equally important is what enters the specification as a consequence. If the financial model needs 0.6 percent degradation to close, the specification must require the module class that delivers it, glass-glass construction, PID resistance tested to IEC 62804, and a linear performance warranty from a manufacturer with a balance sheet likely to outlive the warranty. Modelling optimism is a specification decision wearing a spreadsheet disguise.
Verification after commissioning
Degradation assumptions should not remain assumptions. A capacity test at commissioning establishes the true baseline; annual performance ratio analysis against irradiance data then tracks the actual degradation trajectory. Where the plant includes revenue-grade metering and a pyranometer, this costs almost nothing and converts every future warranty conversation from opinion to evidence. We specify both as standard on any system above 100 kilowatts.
- 01Use 0.8 to 1.2 percent annual degradation for standard modules on the coast, not 0.5.
- 02Model soiling at coastal reality or contract and budget the cleaning regime.
- 03Let the financial model drive the module specification, not the reverse.
- 04Specify a pyranometer and revenue metering; track performance ratio annually.