Load forecasting in electricity distribution

Hourly consumption forecast error fell 28%, reducing balancing costs.

An energy services company produced hourly consumption forecasts for subscriber groups with statistical methods. Ilmek Tahmin generates probabilistic forecasts from weather, calendar and sector signals and uses confidence intervals for balancing.

Result

Forecast error (WAPE) fell 28%; imbalance cost in the balancing market dropped about 17% per year.

Scope

  • Probabilistic forecast (P10/P50/P90)
  • Weather and calendar effects
  • Deviation alerts

Before / after

WAPE-%28
Balancing cost-%17

Similar items

View all — Use Cases

WhatsApp +90 312 555 47 20 merhaba@ilmek.ai