| dc.description.abstract | In keeping with the commitment of a low-emissions society, energy efficiency strategies for the Norwegian urban building stocks shall make a significant contribution to reducing energy use and greenhouse gas emissions.
In the Nordic climate, a large amount of building energy is used for heating purposes. In countries like Norway, where the energy market is dominated by hydro-power, district heating (DH) systems are expected to serve as an alternative heating method to alleviate the increasing pressure on the grid. Furthermore, in the face of green energy initiatives and the increasing share of energy-efficient buildings, there is a pressing need to transform current DH to low-temperature DH (LTDH) to maintain the economic and environmental competitiveness of DH companies in the heating market. The substantially lowered supply temperature of LTDH has broadened the opportunities and challenges to integrating distributed renewable energy resources, requiring enhancement on intelligent heating load prediction.
In the current research on the energy supply systems and building energy demand, in most cases, measured energy data are employed as a package of information, regardless of energy use patterns associated with building types, while most energy forecasts have not yet conducted in-depth studies on sizing or energy demand requirements for typical building types. There lack of a bridge between demand profiles on building stock functions and urban energy supply systems. In addition to the normal condition, Norway and many countries have carried out confinement regulations to hinder the infection spreading in 2020. The distancing measures and changed work regimes have caused significant impacts on energy demand, so it is important to improve the existing knowledge of building operations during unforeseeable disruptions.
To gain a deeper understanding of the energy use and improve the efficiency of Norwegian urban buildings operation, this thesis focused on identifying representative energy trends regarding load profiles and developing appropriate prediction models for Norwegian urban buildings under normal and special conditions. The thesis started with a future development projection of the environmental impacts comparing the buildings with DH and with only electricity, followed by the approach for typical annual energy profiles. Further, a hybrid heating prediction was proposed for sizing and operation. In the end, the energy demand changes due to the COVID- 19 pandemic were studied to examine building operation strategies during special circumstances. Accordingly, four research questions were addressed to fulfill the research goal. In the thesis, the study of building energy use was conducted based on the hourly measured data of kindergartens, schools, nursing homes, and residential buildings. The duration of energy data collection differed by building types and were between two and four years. Since buildings with different floor areas, construction years, and energy labels were involved, to define the representative energy use for buildings with different characteristics, the energy use was first converted into the average specific energy use, W/m2. This applied to the most analysis of research examples in the thesis.
Start with Research Question 1: What are the environmental impacts of heating systems in future building development? It was answered in a study of 28 DHsupplied kindergartens in Norway, where three cases were found depending on the energy share from DH; i.e. DH high share, DH average share, and DH low share. By following different CO2 factors of electricity and local DH production, the typical kindergarten with DH high share had almost the lowest CO2 emission; contrarily, the kindergarten with a lower share of DH or without DH, usually had a wider range of CO2 emissions due to its dependence of the electricity production mix. Then a projection was made by assuming 14.2% growth rate of kindergartens. The result showed that if more than 50-67% of the new building area connected to DH, a smaller increase of CO2 emissions from the projected area could be achieved, depending on the CO2 factors. This proved that buildings with DH were more robust than the ones without DH in terms of CO2 emissions. This top-down question addressed the identification of typical building types for development planning and the necessity for diversifying local energy supply pathways.
Research Question 2: What factors shall be considered for building heating and electricity operation and what are the differences between the two delivered energy forms? It was answered in a study of 40 DH-supplied schools in Norway, using a modified Z-Score to determine working days and holidays, linear regression analysis to predict DH and electricity load profiles, and quality criteria and a cluster method to evaluate the prediction quality. The results showed that the modified ZScores might point out the special energy use periods and show the energy demand trend. Operation of the electric appliances might be concluded with reasonably fast responses by following the attendance, while the DH demand mainly followed the outdoor temperature and the daily work schedule, with a slow control response to short holidays, resulting in a waste of some heat energy. The identified specific load profiles may present the current energy use of schools in the Nordic climate. The predicted annual DH demand was 72 kWh/m2 with a peak load of 48 W/m2; the predicted annual electricity demand was 57 kWh/m2 with a peak load of 18 W/m2. Thus, the buildings with DH may largely reduce the power grid strains. This long-term prediction also highlighted the importance of accurate heating peak load prediction, especially for the promising LTDH, which was addressed in Research Question 3: How can the methods for developing and predicting heating load profiles be improved for future daily LTDH operation? Hereby, a study of 20 DHsupplied nursing homes in Norway proposed a hybrid prediction method, combining long-term DH load prediction by means of linear regression for unit sizing and shortterm (day-ahead) load prediction by means of two Artificial Neural Network models, f72 and g120 (with different input parameters). It was found that including the historical heating loads as the input to the forecasting model improved the prediction quality, especially for the peak load and low-mild heating season, as proved by g120 outperforming in the prediction quality evaluation. Meanwhile, to fulfill the different temperature requirements of domestic hot water and space heating, separate energy conversion units shall be implemented on user-side to upgrade the temperature level of LTDH network.
Lastly, the energy impacts due to COVID-19 were addressed in Research Question 4: What are the energy and economic impacts of the buildings under special circumstances? Since electric heating still accounts for a high share in the country, a study of educational buildings and residential buildings with electric heating was conducted to investigate the lockdown impacts. The results showed that during the 2020 lockdown period, the electricity demand and load profiles for educational buildings were almost the same as in previous years, while there were apparent changes for the residential buildings. Further, three building operation scenarios were proposed: Scenario 1 considered operation under normal settings, Scenario 2 considered the operation of educational buildings under nighttime and weekend settings, and Scenario 3 considered the operation of residential buildings under work-at-home conditions. The scenario-based analysis showed that the electricity demand might be reduced by one-third in educational buildings, between 2.1-4.1 €/(m2·yr) might be saved for kindergartens, and 1.4-2.7 €/(m2·yr) for schools by following Scenario 2. Meanwhile, the electricity density of small apartments varied more significantly than the townhouse. Under Scenario 3, the apartment might spend 2.0-4.1 €/(m2·yr) more for electricity, while the increased bill for the townhouse may be trivial. Moreover, in a community with various building functions, the composition of each building type adopting different working schemes may influence the unit sizing and utilization rate.
To conclude, the proposed methods may be efficiently applied to other public buildings in a similar climate. This allows public authorities to better understand the energy needs of different building functions, project future demand changes taking into account normal situations and future unforeseeable disruptions, and improve the building energy efficiency. | en_US |
