Forecasting of energy balance in green campus
Osipov, Ivan (2018)
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi-fe2018052524776
https://urn.fi/URN:NBN:fi-fe2018052524776
Tiivistelmä
The focus of this thesis is the forecasting of the energy balance. Energy balance can be decomposed to the energy consumption forecasting and energy production forecasting, both of which can help users in making decisions related to the energy storage. It is difficult to store electricity so using forecasting the users can decide if they should plug in more devices or purchase additional electricity.
Time-series data analysis is a well-investigated field, so there are several methods that can be used to implement a predictor of electricity production and consumption. The most commonly used methods of forecasting have been reviewed in this work. Analysis and preprocessing of the input data provided has been developed. Linear Regression and Seasonal Autoregressive Integrated Moving Average and adjusting of settings have been used for the forecasting of electricity consumption and production. The accuracy measurement scores of the forecast have been evaluated.
According to the experiments, different methods are suitable for production and consumption data. This is likely to occur because of the differences in seasonality.
Time-series data analysis is a well-investigated field, so there are several methods that can be used to implement a predictor of electricity production and consumption. The most commonly used methods of forecasting have been reviewed in this work. Analysis and preprocessing of the input data provided has been developed. Linear Regression and Seasonal Autoregressive Integrated Moving Average and adjusting of settings have been used for the forecasting of electricity consumption and production. The accuracy measurement scores of the forecast have been evaluated.
According to the experiments, different methods are suitable for production and consumption data. This is likely to occur because of the differences in seasonality.