Modelling city-scale transient district heat demand by combining physical and data-driven approach

Abstract

A city-scale hourly heat demand model for district heating (DH) is proposed by combining a physical and data-driven approach. The physical heat demand model is based on lumped building parameters of the city determined from easily available data. It enables to investigate the effects from activity, structural, and technology changes on heat demand, e.g. building energy efficiency improvements or climate change induced temperature changes. The RMSQ error was <19% and energy-weighted error <2% for two city cases. The model was applied to Helsinki (60°N) showing 415 GWh year-to-year changes in DH per 1 °C of yearly average ambient temperature change. The year 2050 district heat demand of Helsinki was estimated considering population growth, climate change, and building energy efficiency improvements. Activity-driven factors would increase the heat demand by 17% in 2050 from 2018 in spite of a warmer climate. Through building energy efficiency measures, this increasing trend could be reverted resulting in a 15–19% decrease in the DH demand by 2050, mostly during the peak load. The model enables to analyze changes in temporal demand, which could be useful for integration of renewable energy in cities and communities or for estimating the peak demand often with highest emissions.