Solar Irradiance Calculator

Global Horizontal Irradiance (GHI) is the total solar radiation received on a horizontal surface (W/m² or kWh/m²/day), comprising both direct and diffuse components. Direct Normal Irradiance (DNI) is the beam radiation incident perpendicular to the sun’s rays — critical for concentrating PV and tracking systems. Diffuse Horizontal Irradiance (DHI) is the scattered radiation reaching a horizontal plane from the sky dome, excluding direct beam. These three components are related by: GHI = DNI × cos(θz) + DHI, where θz is the solar zenith angle. IEC 61724-1:2021 mandates their separate measurement for performance ratio validation, and high-quality pyranometers (e.g., ISO 9060:2018 Class A) are required for <2% uncertainty in long-term monitoring.

Tilt angle directly influences POA irradiance by altering the cosine projection of DNI onto the panel surface and modulating diffuse and ground-reflected contributions. For fixed-tilt systems, the optimal annual tilt typically approximates the site’s latitude (±5°), per ASHRAE Guideline 34-2022 and NREL’s PVWatts methodology. However, seasonal optimization may favor steeper tilts in winter (latitude +15°) and shallower in summer (latitude −15°). The calculator uses the Perez transposition model (validated against measured data in Solar Energy Vol. 157, 2017) to compute tilted POA, accounting for anisotropic sky diffusion and horizon brightening — essential for accuracy beyond simple isotropic models.

Azimuth angle determines panel orientation relative to true south (in the Northern Hemisphere) or north (Southern Hemisphere), affecting DNI capture throughout the day. A ±5° azimuth error causes ~0.5–1.2% POA loss near equinoxes at mid-latitudes, rising to >2.5% at high latitudes during solstices — per NREL Technical Report NREL/TP-6A20-80132. East/west deviations disproportionately reduce afternoon/morning yield, impacting load-matching. The calculator applies rigorous solar position algorithms (SPA, based on NOAA’s Reda & Andreas 2003) to resolve incidence angles hourly before integrating daily. For utility-scale projects, IEC 62446-1:2016 requires azimuth verification within ±2° using calibrated compasses or GNSS surveying.

No — this calculator computes monofacial POA irradiance only. Bifacial modules require additional inputs: rear-side albedo (typically 0.1–0.8), module height above ground, row spacing, and ground coverage ratio. Rear irradiance depends on view factor geometry and spectral albedo — addressed in IEC TS 60904-1-2:2021 and the bifacial gain models in PVLib Python (v2.3+). Standard transposition models like Perez underestimate rear yield by 15–40% without explicit bifacial correction. For accurate bifacial energy simulation, pair this tool with dedicated software (e.g., PVSyst v7.4+ or bifacial_radiance) and validate using field measurements per ASTM E3085-21.

Soiling and snow reduce effective POA irradiance by attenuating incident light — but they are *system losses*, not irradiance inputs. This calculator outputs ideal (clean, snow-free) POA; degradation factors must be applied *after* calculation. Typical soiling losses range 2–15% annually (IEC 61724-1:2021 recommends site-specific soiling tables), while snow cover can cause near-total loss for days to weeks. Real-time soiling ratio (SSR) sensors or periodic IV-curve measurements are preferred over fixed derates. Never input reduced GHI/DNI values to simulate soiling — that corrupts transposition physics. Instead, multiply final POA by (1 − soiling_loss) × (1 − snow_loss) as a post-processing step.

Using typical TMY3 or NSRDB data, POA uncertainty ranges 5–12% (k=2), dominated by: (1) GHI/DNI measurement error (±3–5% for Class B pyranometers), (2) transposition model choice (±2–4% vs. measured tilted data), and (3) angular misalignment (±1° tilt/azimuth adds ~0.3–0.8%). Per IEC 61724-1:2021 Annex C, combined standard uncertainty for POA should be reported. High-accuracy applications (e.g., bankability studies) require Class A instruments (ISO 9060:2018) and validated transposition (e.g., SMARTS2 or libRadtran). Ground-truthing with a reference cell on the actual array plane reduces uncertainty to <3%.

Yes — it implements the Perez sky diffuse model (Perez et al., Solar Energy, 1990), which explicitly accounts for circumsolar enhancement (brightening near the sun) and horizon brightening (increased diffuse near the horizon), unlike simpler isotropic or HDKR models. This improves tilted diffuse irradiance accuracy by 5–10% versus isotropic methods, especially at low sun angles and high DNI/DHI ratios. The model uses 14 empirical coefficients fitted to global radiometer data and is recommended in ASHRAE Fundamentals (Ch. 33) and NREL’s SAM documentation. Validation against 12-year BSRN data shows median absolute error <3.5% for POA across tilt angles 0–60° — critical for high-accuracy yield modeling.