Treffer: CLOUD SERVICE FOR ELECTRICITY CONSUMPTION FORECASTING.

Building reliable electricity consumption forecasts is of particular importance for a number of reasons: technological, economic and environmental. The high quality of planning contributes to the formation of an effective operation mode of the Unified Energy System thereby minimizing the consumption of energetic resources and economic costs of forcing the “generation-consumption” equilibrium. The task of the forecasting is computationally complex, and the development of a cloud solution that will assist in its solution seems expedient. This paper discusses the key ideas and principles that formed the basis for the creation of such a software product and also focuses on the details of its practical implementation. The developed cloud service automates the process of planning electricity consumption for subsequent acquisition in a liberalized market. The cloud system is based on a time series subsystem that solves a wide range of tasks. They are storage, manipulation with series, tracking data modifications and much more. A domain-specific language was created to formalize the nature of the relationship between varieties of time series and a description of the calculation rules. The most important component of the cloud solution being developed is the forecasting module, whose principle of operation is to launch a set of predictive models and build the final forecast based on the results of calculations. Prediction models are implemented as stored subroutines in R and Python languages, and the final assumption of the system is carried out either as a competitive selection of the best model based on the retrospective accuracy indicators (MAPE is taken as an estimate) for similar periods, or by weighted voting applied to the models’ results. Mainly developed product allows maintaining a high level of planning quality, providing a final forecast with MAPE within 5%. [ABSTRACT FROM AUTHOR]

Copyright of Proceedings of the International Multidisciplinary Scientific GeoConference SGEM is the property of STEF92 Technology Ltd. and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)