Modelling Influential Factors of Consumption in Buildings Connected to District Heating Systems
Abstract
:1. Introduction
Research Background and Motivation
2. Materials and Methods
2.1. Analysed Dataset
- (1)
- The first data set consists of actual billing data for the final consumers.
- (2)
- The second data set is based on the questionnaires distributed to the apartment owners (tenants).
- (1)
- Subdivision A (SD_A)
- (2)
- Subdivision B (SD_B)
- (3)
- Subdivision C (SD_C)
2.1.1. Billing Data
- (1)
- Customer code
- (2)
- Code of the heat meter
- (3)
- Building area
- (4)
- Building area of the customers with HCAs
- (5)
- Heaters’ thermal power (installed thermal capacity) of the customers with HCAs
- (6)
- Heaters’ thermal power (installed thermal capacity) of the customers without HCAs
- (7)
- Total metered heat
- (8)
- Total number of impulses from all of HCAs in the building
- (9)
- Number of impulses from the specific participating apartment
- (10)
- Floor area of participating apartment
2.1.2. Questionnaire and Interviews
- Information on the age structure and the number of the tenants: this section contained questions about the age structure of the tenants in years prior to and after the installation of HCAs. This information is considered to be important with the presumption that certain age groups (specifically, children under the age of 7 and retirees) spend scientifically more time in their apartments then school children, students and people in work. Additionally, the presumption is that the desired heat comfort level for children under the age of 7 years is higher than that for other groups. The answer to this question is asked for the years 2010–2017, while 2015 was the year HCA were installed in this building. The categories of age groups have been defined as:
- ○
- Ages 0–6 years
- ○
- Ages 7–18 years
- ○
- Ages 19–65 years
- ○
- Ages 65 and more
- The information of the floor is asked for the purposes of assessing the influence on the consumption of the location of the specific apartment within the building.
- Information on the daily occupancy time: this section is grouped within three groups as shown in Table 2.
- Information on the existence of the efficient windows.
- Information on the type of the windows: in this section participants needed to quality their window types in predefined categories. Predefined categories are chosen based on the knowledge of the most usual windows type in Croatia, as shown in Table 2.
- Preferred heat comfort level: assessed in three categories.
- Number of unheated rooms: historically in this geographic area there is a custom not to heat the rooms person/family is not using during the daytime.
- Average daily ventilation rate: the last question is related to the influence of the ventilation rate of the apartment on the consumption and it is presumed to be a behavioural parameter of a significant influence.
2.2. Statistical Analysis
2.2.1. Frequency Density of Qualitative Variable
2.2.2. Frequency Density of Qualitative Variable
- (1)
- Mean: For arithmetic mean a measure of central tendency.
- (2)
- Standard Deviation: Measure the amount of variation using the square root of the variance and gives an information how close are all of occurrences clustered around the mean of a particular dataset.
- (3)
- Minimum Value: Gives a minimum value in a dataset.
- (4)
- Q1 Value: Gives a middle value between the minimum value and median in a dataset.
- (5)
- Median: The measure the central tendency giving the middle value in dataset.
- (6)
- Q3 Value: Gives a middle value between the median and the maximum value in a dataset.
- (7)
- Maximum Value: Gives a maximum value in a dataset.
- (8)
- Median Absolute Deviation: A scale estimate based on the median absolute deviation.
- (9)
- Interquartile range (IQR): A measure of the spread between the 1st and 3rd quartile.
- (10)
- Coefficient of variation or relative standard deviation (CV): Represents a ration between standard deviation and a mean, or a spread relative to its expected value.
- (11)
- Skewness: Measure of the symmetry of the distribution. Positive skewness indicates a tail on the right side of the graph, while negative skewness indicates a tail on a left side of the graph. A value of skewness of zero indicates a symmetric distribution or an asymmetric distribution when “fat and short tail” is balancing the “long and thin tail”.
- (12)
- Standard Error of Skewness: Standard deviation or a sampling distribution of skewness.
- (13)
- Kurtosis: Describes the shape of the distribution or the outlook of tails in the distribution, describing the tails based on the kurtosis as either heavy or light. Positive kurtosis is called leptokurtic and indicates a distribution with heavier tails than a normal distribution. Negative kurtosis is called platykurtic and indicates a distribution with lighter tails than a normal distribution.
2.2.3. Frequency Density of Qualitative Variable
2.2.4. Frequency Density of Qualitative Variable
- (1)
- Monthly number of impulses from HCAs in an individual apartment (from Billing Data, denoted—ID_Imp)
- (2)
- Monthly total number of all HCA impulses in a building (from Billing Data, denoted—B_Imp)
- (3)
- Total metered heat on a main building heat meter (from Billing Data, denoted—B_E_MWh)
- (4)
- Heated floor area of an individual apartment (from Billing Data, denoted—m2)
- (5)
- Monthly absolute allocated heat to an individual apartment (from Billing Data, denoted—Ap_MWh)
- (6)
- Monthly specific heat consumption in kWh per m2 (from Billing Data, denoted—Ap_kWh_m2)
- (7)
- Specific monthly ratio of total impulses from HCAs of an individual apartment relative to heated floor area (from Billing Data, denoted—Imp_m2)
- (8)
- Total number of dwellers in an individual apartment (from Billing Data, denoted—Dwellers)
- (1)
- Floor (from both Billing Data and Questionnaire Data, denoted Floor)
- (2)
- Average daily occupancy of an individual apartment (from Questionnaire Data, denoted AverageDailyOccupancy)
- (3)
- Data on efficiency of apartment’s windows (from Questionnaire Data, denoted EfficientWindows)
- (4)
- Data on type of apartment’s windows (from Questionnaire Data, denoted EffWindowType)
- (5)
- Data on the desired heat comfort level (from Questionnaire Data, denoted ThermalComfort)
- (6)
- Data on number of unheated rooms (from Questionnaire Data, denoted UnheatedRooms)
- (7)
- Data on estimated ventilation rate (from Questionnaire Data, denoted Ventilation)
- (8)
- Total number of dwellers in an individual apartment (from Billing Data, denoted—Dwellers)
3. Results
3.1. Normality of Data
3.1.1. Shapiro-Wilk Normality Test
3.1.2. Visualising the Fit
3.2. Frequencies of Qualitative Parameters from Questionnaires
3.3. Descriptive Statistics
3.3.1. Quantitative Predictors
3.3.2. Correlations and Relationships of Quantitative Variables
- −1 indicates a strong negative correlation (can be interpreted as variable A increases. variable A decreases)
- 0 indicates no correlation between the pair of variables. and.
- 1 indicates strong positive correlation (can be interpreted as variable A increases. variable B increases as well).
3.3.3. Visualising the Sample Distributions
3.3.4. Qualitative Predictors
- (1)
- Total number of dwellers in an individual apartment. The largest spread of values is found, as expected, in the apartment with the largest number of dwellers.
- (2)
- Floor. The largest specific consumption was recorded for the basement apartments, as expected.
- (3)
- Average daily occupancy of an individual apartment. No distinguished influence of the daily occupancy mode on the heat consumption can be interpreted. It can be concluded that the apartments in which dwellers stay most of the day can be expected to have the largest spread of values for the specific consumption.
- (4)
- Efficiency of apartment’s windows. As expected, apartments without efficient windows have larger heat consumption.
- (5)
- Type of apartment’s windows. Type of windows are presented.
- (6)
- Desired heat comfort level. As expected, apartments with requirements for warmer ambient temperature have larger consumption. None of the apartments that have participated in survey have stated they preferred colder thermal comfort level.
- (7)
- Number of unheated rooms. No distinguished influence of the daily occupancy on the specific consumption can be interpreted.
- (8)
- Ventilation rate. As expected, larger specific heat consumption is recorded in apartments with larger frequency of ventilation of space by opening the windows, while the mean for all three groups (often, occasionally and rarely) is at a similar level.
3.4. Multiple Linear Regression Analysis
- The most influential factor among quantitative variables are the number of impulses in the apartment. allocated and measured heat for both the apartment and the building, and the ratio between allocated impulses and the area (the most significant factor).
- Among the quantitative variables the largest influence is recorded from the particular window type, following the daily occupancy rate and frequency of ventilation.
- The adjuster R2 for the analysed model is 0.998, meaning that 99.8% of analysed data can be explained by the variables obtained in Table 5.
- If we were to express the function using Equation (5), the “Estimates” in the Table 5 would be equivalent to the regression coefficients (bi), while the predictors would have a Xi value.
4. Discussion
Supplementary Materials
Supplementary File 1Funding
Conflicts of Interest
Abbreviations
IEA | International Energy Agency |
EU | European Union |
DHW | Domestic Hot Water |
HCA | Heat Cost Allocators |
DHS | District Heating Systems |
SD-A | sub-division A of the analysed building |
SD-B | sub-division B of the analysed building |
SD-C | sub-division C of the analysed building |
HRK | Croatian Kuna, currency |
IQR | Interquartile Range |
CV | Coefficient of Variation |
References
- Fraunhofer and Alia. Mapping and Analyses of the Current and Future (2020–2030) Heating/Cooling Fuel Deployment (Fossil/Renewables); ENER/C2/2014-641; Fraunhofer and Alia: Berlin, Germany, 2015. [Google Scholar]
- International Energy Agency. Energy Efficiency Market Report 2015; International Energy Agency: Paris, France, 2015. [Google Scholar]
- de Boeck, L.; Verbeke, S.; Audenaert, A.; de Mesmaeker, L. Improving the energy performance of residential buildings: A literature review. Renew. Sustain. Energy Rev. 2015, 52, 960–975. [Google Scholar] [CrossRef]
- Catalina, T.; Virgone, J.; Blanco, E. Development and validation of regression models to predict monthly heating demand for residential buildings. Energy Build. 2008, 40, 1825–1832. [Google Scholar] [CrossRef]
- Haghighat, L.M.F. Multiobjective optimization of building design using TRNSYS simulations genetic algorithm and Artificial Neural Network. Build. Environ. 2010, 45, 739–746. [Google Scholar]
- Al-Ajmi, F.F.; Hanby, V.I. Simulation of energy consumption for Kuwaiti domestic buildings. Energy Build. 2008, 40, 1101–1109. [Google Scholar] [CrossRef]
- International Energy Agency. Mapping the Energy Future: Energy Modelling and Climate Change Policy; Energy and Environment Policy Analysis Series; International Energy Agency: Paris, France, 1998. [Google Scholar]
- Kavgic, M.; Mavrogianni, A.; Mumovic, D.; Summerfield, A.; Stevanovic, Z.; Djurovic-Petrovic, M. A review of bottom-up building stock models for energy consumption in the residential sector. Build. Environ. 2010, 45, 1683–1697. [Google Scholar] [CrossRef]
- Yang, L.; Yan, H.; Lam, J.C. Thermal comfort and building energy consumption implications—A review. Appl. Energy 2014, 115, 164–173. [Google Scholar] [CrossRef]
- Nguyen, T.A.; Aiello, M. Energy intelligent buildings based on user activity: A survey. Energy Build. 2013, 56, 244–257. [Google Scholar] [CrossRef] [Green Version]
- Olofsson, T.; Andersson, S.; Sjogren, J.U. Building energy parameter investigations based on multivariate analysis. Energy Build. 2009, 41, 71–80. [Google Scholar] [CrossRef]
- Jain, R.K.; Smith, K.M.; Culligan, P.J.; Taylor, J.E. Forecasting energy consumption of multi-family residential buildings using support vector regression: Investigating the impact of temporal and spatial monitoring granularity on performance accuracy. Appl. Energy 2014, 123, 168–178. [Google Scholar] [CrossRef]
- Smola, A.J.; Vishwanathan, S.V.N.; Le, Q.V. Bundle Methods for Machine Learning. In Advances in Neural Information Processing Systems 20; Platt, J.C., Koller, D., Singer, Y., Roweis, S., Eds.; MIT Press: Cambridge, MA, USA, 2008. [Google Scholar]
- Lund, H.; Möller, B.; Mathiesen, B.V.; Dyrelund, A. The role of district heating in future renewable energy systems. Energy 2010, 35, 1381–1390. [Google Scholar] [CrossRef]
- Thellufsen, J.Z.; Lund, H. Energy saving synergies in national energy systems. Energy Convers. Manag. 2015, 103, 259–265. [Google Scholar] [CrossRef]
- Commision to the European Parlament the Council. The European Economic and Social Committee and the Committee of the Regions. An EU Strategy on Heating and Cooling; Commision to the European Parlament the Council: Brussels, Belgium, 2016. [Google Scholar]
- European Commission. Directive 2012/27/Eu of the European Parliament and of the Council of 25 October 2012 on Energy Efficiency. Amending Directives 2009/125/EC and 2010/30/EU and repealing Directives 2004/8/EC and 2006/32/EC. 2012. Available online: https://www.energypoverty.eu/publication/directive-201227eu-european-parliament-and-council-25-october-2012-energy-efficiency (accessed on 7 January 2019).
- Siggelsten, S. Reallocation of heating costs due to heat transfer between adjacent apartments. Energy Build. 2014, 75, 256–263. [Google Scholar] [CrossRef]
- Cholewa, T.; Siuta-Olcha, A. Long term experimental evaluation of the influence of heat cost allocators on energy consumption in a multifamily building. Energy Build. 2015, 104, 122–130. [Google Scholar] [CrossRef]
- Carpino, C.; Mora, D.; de Simone, M. On the use of questionnaire in residential buildings. A review of collected data methodologies and objectives. Energy Build. 2019, 186, 297–318. [Google Scholar] [CrossRef]
- Andersen, R.V.; Toftu, J.; Andersen, K.K.; Olesen, B.W. Survey of occupant behaviour and control of indoor environment in Danish dwellings. Energy Build. 2009, 41, 11–16. [Google Scholar] [CrossRef]
- Santin, O.G. Behavioural Patterns and User Profiles related to energy consumption for heating. Energy Build. 2011, 43, 2662–2672. [Google Scholar] [CrossRef]
- Sardianou, E. Estimating space heating determinants: An analysis of Greek households. Energy Build. 2008, 40, 1084–1093. [Google Scholar] [CrossRef]
- Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on the Protection of Natural Persons with Regard to the Processing of Personal Data and on the Free Movement of Such Data. And Repealing Directive 95/46/EC (General Data Protection Regulation). Available online: https://www.lewik.org/term/13554/regulation-eu-2016-679-of-the-european-parliament-and-of-the-council-of-27-april-2016-on-the-protection-of-natural-persons-with-regard-to-the-processing-of-personal-data-and-on-the-free-movement-of-such-data-and-repealing-directive-95-46-ec-general-data-protection-regulation-gdpr/ (accessed on 7 January 2019).
- Regulation on the Method of Distribution and Calculation of Costs for the Delivered Heat Energy (Official Gazzette. br. 99/14. 27/15. 124/15). Available online: https://narodne-novine.nn.hr/clanci/sluzbeni/2014_08_99_1956.html (accessed on 7 January 2019).
- R Core Team. R: A Language and Environment for Statistical Computing; R. Foundation for Statistical Computing: Vienna, Austria, 2018; Available online: https://www.R-project.org/ (accessed on 7 January 2019).
- Juodis, E.; Jaraminiene, E.; Dudkiewicz, E. Inherent variability of heat consumption in residential buildings. Energy Build. 2009, 41, 1188–1194. [Google Scholar] [CrossRef]
Characteristics | SD_A | SD_B | SD_C |
---|---|---|---|
Total floor area | 3383.18 | 3500.04 | 1880.53 |
Total number of apartments | 61 | 61 | 32 |
Total number of apartments with HCA | 58 | 61 | 31 |
Floor area of apartments with HCA | 3246.79 | 3500.04 | 1793.56 m2 |
Floor area of apartments without HCA | 136.39 | 0 | 86.97 m2 |
Thermal capacity of apartments with HCA | 410.731 | 394.260 | 227.389 kW |
Thermal capacity of apartments without HCA | 17.254 | 0 | 11.026 kW |
Total number of tenants | 111 | 128 | 70 |
Number of floors | 6 | 6 | 5 |
|
Dataset for Year | p |
---|---|
2016 | 0.2151 |
2017 | 0.0975 |
Value | ID_Imp | B_Imp | B_E_MWh | m2 | Dwellers | Ap_MWh | Ap_kWh_m2 | Imp_m2 |
---|---|---|---|---|---|---|---|---|
Mean | 548.74 | 28,901.33 | 38.27 | 57.66 | 2.09 | 0.69 | 11.88 | 9.39 |
.Stand. Deviation | 545.79 | 19,716.36 | 24.69 | 20.4 | 1.25 | 0.65 | 9.54 | 8.04 |
Min | 0 | 2140 | 3.05 | 29.72 | 0 | 0 | 0.13 | 0 |
Q1 | 133 | 12,720 | 21.71 | 35.3 | 1 | 0.19 | 3.68 | 2.5 |
Median | 406 | 24,663 | 32.94 | 55.37 | 2 | 0.54 | 10.27 | 7.77 |
Q3 | 781 | 41037 | 51 | 73.98 | 3 | 0.97 | 17.75 | 14.24 |
Max | 3444 | 82,382 | 95.91 | 92.99 | 5 | 3.89 | 50.16 | 45.26 |
MAD | 453.68 | 18,925.39 | 20.87 | 27.89 | 1.48 | 0.56 | 10.26 | 8.55 |
IQR | 648 | 28,317 | 29.29 | 38.68 | 2 | 0.78 | 14.07 | 11.74 |
CV | 0.99 | 0.68 | 0.65 | 0.35 | 0.6 | 0.94 | 0.8 | 0.86 |
Skewness | 1.6 | 0.71 | 0.66 | 0.16 | 0.75 | 1.55 | 0.87 | 0.91 |
SE. Skewness | 0.05 | 0.05 | 0.05 | 0.05 | 0.05 | 0.05 | 0.05 | 0.05 |
Kurtosis | 3.04 | −0.14 | −0.26 | −1.35 | −0.36 | 2.83 | 0.34 | 0.44 |
Number. Valid | 2465 | 2465 | 2465 | 2465 | 2465 | 2465 | 2465 | 2465 |
Percentage. Valid | 99.96 | 99.96 | 99.96 | 99.96 | 99.96 | 99.96 | 99.96 | 99.96 |
Predictors | Estimate | Standardised (βi) | Std. Error | p | Sign. Code 1 |
---|---|---|---|---|---|
(Intercept) | 0.833 | 0.000 | 0.167 | <0.001 | *** |
No of impulses for an apartment | −0.014 | −0.819 | <0.001 | <0.001 | *** |
No of impulses for a building | <0.001 | −0.207 | <0.001 | <0.001 | *** |
Total heat for a building in MWh | 0.086 | 0.218 | 0.006 | <0.001 | *** |
Heated floor area of an apartment | −0.008 | −0.018 | 0.001 | <0.001 | *** |
Number of tenants | 0.006 | 0.001 | 0.016 | 0.722 | - |
Allocated MWh per apartment | 11.730 | 0.831 | 0.270 | <0.001 | *** |
Ratio between number of impulses and area | 1.166 | 0.985 | 0.005 | <0.001 | *** |
Daily occupancy—rarely | −0.124 | −0.005 | 0.039 | 0.001 | ** |
Daily occupancy—whole day | −0.049 | −0.002 | 0.048 | 0.305 | - |
Efficient Windows—partly | −0.093 | −0.002 | 0.140 | 0.508 | - |
Efficient Windows—yes | −0.179 | −0.005 | 0.122 | 0.141 | - |
Windows Type—Plastic Double Glass | −0.178 | −0.009 | 0.099 | 0.073 | . |
Windows Type—Wooden Double Glass | −0.064 | −0.002 | 0.113 | 0.575 | - |
Windows Type—No Efficient Windows | −0.074 | −0.004 | 0.101 | 0.464 | - |
Windows Type—Plastic | 0.026 | 0.001 | 0.119 | 0.825 | - |
Thermal Comfort—Warmer | 0.030 | 0.002 | 0.034 | 0.379 | - |
Unheated Rooms—One | 0.001 | <0.001 | 0.040 | 0.983 | - |
Unheated Rooms—Three | −0.256 | −0.004 | 0.119 | 0.032 | * |
Unheated Rooms—Two | −0.115 | −0.005 | 0.048 | 0.018 | * |
Ventilation—Often | −0.077 | −0.003 | 0.037 | 0.036 | * |
Ventilation—Rarely | 0.133 | 0.003 | 0.078 | 0.090 | . |
Observations | 836 | - | - | - | - |
R2 /adjusted R2 | 0.998/0.998 | - | - | - |
© 2019 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
Share and Cite
Maljkovic, D. Modelling Influential Factors of Consumption in Buildings Connected to District Heating Systems. Energies 2019, 12, 586. https://doi.org/10.3390/en12040586
Maljkovic D. Modelling Influential Factors of Consumption in Buildings Connected to District Heating Systems. Energies. 2019; 12(4):586. https://doi.org/10.3390/en12040586
Chicago/Turabian StyleMaljkovic, Danica. 2019. "Modelling Influential Factors of Consumption in Buildings Connected to District Heating Systems" Energies 12, no. 4: 586. https://doi.org/10.3390/en12040586
APA StyleMaljkovic, D. (2019). Modelling Influential Factors of Consumption in Buildings Connected to District Heating Systems. Energies, 12(4), 586. https://doi.org/10.3390/en12040586