Model for forecasting residential heat demand based on natural gas consumption and energy performance indicators

The forecasting of energy and natural gas consumption is a topic that spans different temporal and spatial scales and addresses scenarios that vary significantly in consistency and extension. Therefore, although forecasting models share common aims, the specific scale at which each model has been developed strongly impacts its features and the parameters that are to be considered or neglected. There are models designed to handle time scales, such as decades, years, and months, down to daily or hourly models of consumption. Similarly, there are patterns of forecasted consumption that range from continents or groups of nations down to the most limited targets of single individual users, passing through all intermediate levels. This paper describes a model that is able to provide a short-term profile of the hourly heat demand of end-users of a District Heating Network (DHN). The simulator uses the hourly natural gas consumptions of large groups of users and their correlation with the outside air temperature. Next, a procedure based on standards for estimating the energy performance of buildings is defined to scale results down to single-user consumption. The main objective of this work is to provide a simple and fast tool that can be used as a component of wider models of DHNs to improve the control strategies and the management of load variations. The novelty of this work lies in the development of a plain algebraic model for predicting hourly heat demand based only on average daily temperature and historical data of natural gas consumption. Whereas aggregated data of natural gas consumption for groups of end users are measured hourly or even more frequently, the thermal demand is typically evaluated over a significantly longer time horizon, such as a month or more. Therefore, the hourly profile of a single user’s thermal demand is commonly unknown, and only long-term averaged values are available and predictable. With this model, used in conjunction with common weather forecasting services that reliably provide the average temperature of the following day, it is possible to predict the expected hourly heat demand one day in advance and day-by-day.