As the energy transition accelerates, wind power stands tall as one of the most vital sources of renewable energy. It’s a critical driver of decarbonization, but like solar, the intermittency of its renewable resource, wind, presents a challenge for asset owners and operators. Forecasting wind generation requires a complex understanding of numerous weather variables and their influence on turbine operations.
At Amperon, we’re excited to announce our latest product expansion in asset renewables – Asset Wind Short-term and Sub-hourly Forecasts – a solution that combines traditional and new modern forecasting methodologies to improve accuracy and maximize asset efficiency for utilities and independent power producers. Our short-term forecast provides hourly updates up to 15 days and our sub-hourly forecast provides granularity up to 5 minutes up to 15 days.
Wind asset owners rely on accurate and granular (sub-hourly) forecasts to manage crucial operations including balancing net demand as load serving entities (LSEs) and maximizing financial efficiency of assets in day-ahead and real-time markets. Wind asset owners bidding their generation in wholesale markets are subject to financial losses if they over-schedule or under-schedule in both day-ahead and real-time markets. Especially with European TSOs moving to 15- and 30-minute granularity for day-ahead markets, accurate and sub-hourly wind forecasts are mission-critical to confidently scheduling bids.
Wind generation forecasting is notoriously complex, typically with wider accuracy bands than solar forecasting due to greater variability in weather forecasts and limitations in physics-based approaches. Similar to our innovative approach to solar generation forecasting, we are addressing similar forecasting challenges in wind. Our improvements are driven by unique features and modeling techniques, including:
Wind resource forecasts are an essential input into wind generation forecasts. The accuracy of variables like wind speed and wind gust drastically impacts generation output. This uncertainty is magnified if the generation forecast is relying on a single third-party or in-house weather source. Amperon addresses this weather error risk by intelligently applying multiple weather vendor forecasting sources into a dynamically weighted ensemble weather model. This approach evaluates the accuracy of multiple weather vendors on an hourly basis and calibrates their weighting as part of an ensemble weather model used in our final generation prediction, minimizing the risk of errors introduced from any single provider. This feature is available in our hourly short-term forecast product for both asset-level solar and wind.
Amperon enables customers to upload historic curtailment, potential power, and plant availability data as part of onboarding and on-going data updates. This additional data provides necessary context to identify causes of underperformance (i.e., lower-than-expected output). Specifically, it improves our ML models by ensuring they can differentiate performance issues due to weather-related factors (i.e., low wind speeds) from exogenous factors like utility-imposed curtailment or scheduled maintenance.
As wholesale electricity prices in Europe and North America become increasingly volatile, wind asset owners must depend on accurate and granular generation forecasts to optimize their assets’ financial performance.
We are committed to continuous improvement of our products and forecasting models. We would love to hear from you. Contact us today to discuss ways Amperon can help improve your wind forecasting accuracy.
