The Outage Prediction Model uses a machine-learning, predictive model to combine historical weather data and up-to-date, hyperlocal weather forecasts with historical outage data for the service area. This combination allows the utility to see what areas may be hit hardest and decide where to pre-stage restoration crews and equipment.
The prediction model allows utilities to:
- · Review an automated storm prediction with 72-hour lead time as weather models update in real time.
- · Perform historical storm searches, comparing current weather forecast conditions and predicted outages to past outcomes under similar weather conditions.
- · Plan for best and worst-case scenarios looking at variable factors, such as wind speed, precipitation, temperature and humidity.
The prediction model accounts for parameters, including atmospheric pressure, soil moisture, and foliage. It includes other data such as types of vegetation, seasonality, or other information such as assets that could be incorporated based on location, weather regime and system design. The model allows for an individualized solution for each situation.