Traditional forecasting models often fail during periods of significant change. Infrastructure operators for natural gas, in particular, are facing such changes. Sales forecasts with new procurement scenarios, assessments of balancing energy or fuel gas requirements, predictions of capacity utilisation, and the scaling up of the hydrogen core network - all these require advanced forecasting based on artificial intelligence (AI).
By connecting our cloud-based forecasting service cpX.AI, the transmission system operator ONTRAS can use continuously predicted measured values for various market processes in the balancing and procurement process or as intelligent substitute values in the event of measurement data failures.
The forecasting service combines simple statistical forecasting models with neural-based networks (machine learning).
In addition to external factors such as weather data, holiday calendars, price indices, etc., historical measured values form the basis for the cyclical training of the forecast models. For this purpose, cpX.AI was directly connected to the ONTRAS energy data management system. Monitoring in cpX.AI supports ONTRAS in displaying completeness, anomalies, forecast quality and forecast errors.
ONTRAS operates a 7,700-kilometer long transmission system in eastern Germany and ensures the reliable and efficient transport of gaseous energy - today and tomorrow. ONTRAS is relentlessly driving forward sustainable technical and network planning solutions for the integration and transport of climate-neutral gases, particularly hydrogen.
The following screenshot shows a comparison of a measurement with the forecast for a defined forecast object:
