Aiolos Forecast Studio Electricity Load Forecasts Manage all your forecasts in Aiolos - the user friendly forecasting software PHONE +46 90 15 49 00 info.energy@vitec.se www.vitec.se
about aiolos Aiolos is the market leading tool for effective management of all your consumption and production forecasts. With Aiolos it is possible to predict the electricity load, district heating load, district cooling load, wind power production, solar power production, gas consumption and hydro power with advanced and reliable forecast models. Short-term and mid-range forecasts with a resolution of 5 minutes to 1 hour are available as well as year long forecasts based on statistical weather series (normal weather). Aiolos can also be used as a framework to collect, view and compare forecasts made in other systems. The system makes several forecasts based on different weather forecasts, which then are weighted together into a final forecast. The Aiolos electricity load forecasting module is used by the majority of the TSO s and energy traders in the Nordic region. The system makes short-term forecasts but also supports year-long forecasts used for financial trading. The combination of advanced, predictive modeling and a modern and user-friendly interface, makes Aiolos a musthave tool in today s energy trading. aiolos forecasts: electricity load district heating load gas consumption wind power generation solar energy production hydro power production Aiolos Forecast Studio
users For example Axpo Dala Kraft Elektra Invest Energiakolmio E.ON Fingrid Fortum Markets Göteborg Energi DinEl Hafslund Helsinki Energy Litgrid Nordic Power Management Skellefteå Kraft Statkraft Svenska Kraftnät Vattenfall
set up Define the series you want to forecast, which weather data to use and how to aggregate the series. A load series can be the load for one city/region or an aggregated series like the consumption/production of a whole country. forecasts You can fast and easily make forecasts and view them alongside with the historical data for the chosen forecast series. If you have many series to forecast you can choose the series that you want to forecast by a manual check, a template or by the search function.
validation Compare the load forecasts with the outcome of other days, for example same day types or days with a similar weather situation. View the load forecast together with the outcome of other days in the graph and see which weather data that was used for those days you are comparing. The graphical interface makes it easy to analyze if the weather data can explain the difference in the load forecast. follow up Aiolos has a very powerful and user friendly module for evaluating the forecasts made in the system. You can evaluate the quality of the forecasts on different time horizons, which weather suppliers that make the best load/production forecasts, which combination of weather suppliers that makes the best load/production forecasts, which forecasting models within Aiolos that makes the best forecasts as well as comparing different forecasting suppliers with each other. The result can be displayed in MAE%, RMSE% etc.
one studio for all your forecasts Aiolos Forecasts Studio is built to be your tool for managing all of your company s forecasts, regardless of in what system they were made. It is easy to import them to Aiolos Forecast Studio and once they are there, you can compare them with each other, weigh them together and view them together in an user friendly way. Hydro Choose the series you want to forecast individually or use a template Electricity Solar Wind Power Compare Aiolos with other systems Are you using another forecasting systems today? If so, keep doing that but import those load forecasts to Aiolos. Then with Aiolos you can: - Compare different weather forecast suppliers - Compare different forecast systems - Weigh different systems/suppliers together - View different forecasts in the same diagram - Analyze how different suppliers performs in terms of forecast horizons
The red line separates the forecast from the historical data All data can be viewed either in the worksheet or in the graph The weather data used for the forecast is shown in the graph
model description 1. The model 2. Input 3. Load series Model load history Aiolos uses time series and regression to find the correlations between the social-, weather- and stochastic parts of the load. Input to the model are historical load data, historical weather data and weather forecasts. In Aiolos you define load you are making forecast series is a time series wi defined geographical ar entity such as a compan 6. The social dependence 7. The weather dependence 8. The stochastic lo Day types are defined with a day type calendar where the different days have a specific load pattern. Each load series can have its own calendar. The day patterns are automatically found by Aiolos. Aiolos takes in to consideration the dependence of temperature, global radiation and wind speed. The dependence is found using stepwise linear regression on the historical load and the historical weather. The stochastic load is that cannot be explain or weather dependenc crises, special events e
4. Aggregated series 5. Parameters P(t) = Ps(t) + Pw(t) + σ(t) series, i.e. series s for. A load th load data for a ea, or some other y. The user defines how the load series are to be aggregate. Mathematical operation can be applied to the aggregations. The total load for a load series is based on day type patterns, weather parameters and a stochastic load. Every load series is calculated separately with regard to its individual characteristics. ad 9. Adaption 10. The final forecast that part of the error ed with the sociale. E.g. financial tc. By comparing historical forecasts with the actual outcome of the load the system adjusts the load forecast to the current situation. The final forecast is done by the click of a button, and is now ready to be analyzed. The final forecast is accomplished in a matter of minutes, by a single click, even though it involves these advanced calculations for maybe thousands of series.
key features Multiple weather forecasts Aiolos uses several weather forecasts (from several weather suppliers) when making its load forecasts. This means that for each forecast series, the system actually creates up to 5 load forecasts, made by different weather data and there after weighted together in an optimal way. By using multiple weather forecasts, the accuracy of the forecasts improves substantially. Autotuning With the Autotune module, Aiolos helps the user evaluate what settings you should use to create the best load forecasts. For example, how many years of historical data generates the best forecasts? Follow up Use Aiolos to create the reports that you are interested in, with just a mouse click you can get a reports for: - The load forecast accuracy - The accuracy of the load forecast made by different weather suppliers Real-time data If you have real time measurements for the consumption/production, that can be used when making the forecasts and will improve the really short forecasts, up to 6-8 hours. Those forecasts can be used for the intraday trading. Multiple models Aiolos uses different models for different types of customers. Choose the forecast model that is best suited for the specific customer or let the system optimize that setting for you. Flexible resolution Aiolos is built to handle resolutions from 5 minutes and up to 1 hour. Dynamic fractions By adding a regression on the forecasted series, the forecast quality can be improved even more. This can also involve forecasts made by external models. - The accuracy of the load forecast made by different forecasting models - Analyze which combination of weather suppliers that makes the best load forecasts depending on seasons of the year.
aiolos windpower Aiolos forecasts wind power for everything from individual wind turbines, to wind farms, and up to an entire country s production. The system bases its forecasts on a hybrid model. By combining a statistical model with physical factors such as wind speed at different altitudes, wind direction, air density and so on, Aiolos produces production forecast of very high quality. Aiolos WindPower uses up to 5 different weather suppliers when making the production forecasts. This means that the system for each wind turbine/park creates 5 production forecasts which are then weighted together to the final and optimal production forecast. Aiolos WindPower can be integrated in Aiolos Forecast Studio or be used as an stand-alone software.
do you want a demo?...or do you want to test the system? The only way to know for certain how accurate forecasts Aiolos can make for your customers, is to test the system for a few months. It s very easy to set up the system for a test and you can evaluate the system without installing anything on your servers. Contact us for more information about a test. Contact information Today, about 50 energy traders are using Aiolos for their daily forecasting. Our sales team would be more than happy to visit you and your company to see if Aiolos could be a forecasting tool for you. If Aiolos sounds interesting, do not hesitate to contact us for a demo. Please contact Niklas Berg at +46 706 186 490 or niklas.berg@vitec.se Vitec Energy AB PHONE +46 90 15 49 00 info.energy@vitec.se www.vitec.se