Dashboard
Real-time solar generation forecast for utility-scale PV plants in India · pvlib physics + Open-Meteo live weather + trained ML model.
Open-Meteo· live weatherpvlib· physicsML model· solar_forecast_model.pkl
Formulas used
Irradiance closureOFFICIAL_SOURCE · pvlib
GHI = DNI · cos(θz) + DHI
GHI — Global Horizontal Irradiance (W/m²)
DNI — Direct Normal Irradiance (W/m²)
θz — Solar zenith angle (sun-to-vertical)
DHI — Diffuse Horizontal Irradiance (W/m²)
PV generation (PVWatts)OFFICIAL_SOURCE · pvlib
DC = pvwatts_dc(poa, t_cell, pdc0, γ=-0.0035) AC = pvwatts_inverter(DC, η=0.96) MW = min(AC / 1e6, capacity_mw)
poa — Plane-of-array irradiance (W/m²)
t_cell — Cell temperature from Faiman model (°C)
pdc0 — Nameplate DC capacity (W)
γ — Temperature coefficient (−0.0035 /°C, c-Si default)
DSM deviation (interval-normalized)USER_CONFIGURABLE
deviation_pct = (|actual − scheduled| / Δt_hours)
× block_hours / denominator × 100Δt_hours — Evaluation interval in hours
block_hours — DSM time-block length (profile.time_block_minutes / 60)
denominator — available_capacity (CERC) or scheduled (simple mode)
Confidence scoreFALLBACK_DEFAULT
confidence = clamp(1 − 0.35 · cloud_fraction, 0.4, 0.99)
cloud_fraction — Cloud cover fraction (0–1) from weather data
Provenance: REAL_BENGALURU + MODEL_LEARNED — Live weather from Open-Meteo at real coordinates; generation derived via pvlib physics and the trained ML irradiance model.
Forecast Parameters
Quick presets: