Analysis and clustering of natural gas consumption data for thermal energy use forecasting

In this paper, after a brief analysis of the connections between the uses of natural gas and thermal energy use, the natural gas consumption data related to Italian market are analyzed and opportunely clustered in order to compute the typical consumption profile in different days of the week in different seasons and for the different class of users: residential, tertiary and industrial. The analysis of the data shows that natural gas consumption profile is mainly related to seasonality pattern and to the weather conditions (outside temperature, humidity and wind chiller). There is also an important daily pattern related to industrial and civil sector that, at a lower degree than the previous one, does affect the consumption profile and have to be taken into account for defining an effective short and mid term thermal energy forecasting method. A possible mathematical structure of the natural gas consumption profile is provided. Due to the strong link between thermal energy use and natural gas consumption, this analysis could be considered the first step for the development of a model for thermal energy forecasting.