Linear and Neural Network-based Models for Short-Term Heat Load Forecasting

911 Views
611 Downloads
Export citation: ABNT
POTOČNIK, Primož ;STRMČNIK, Ervin ;GOVEKAR, Edvard .
Linear and Neural Network-based Models for Short-Term Heat Load Forecasting. 
Strojniški vestnik - Journal of Mechanical Engineering, [S.l.], v. 61, n.9, p. 543-550, june 2018. 
ISSN 0039-2480.
Available at: <https://www.sv-jme.eu/article/linear-and-neural-network-based-models-for-short-term-heat-load-forecasting/>. Date accessed: 01 jun. 2020. 
doi:http://dx.doi.org/10.5545/sv-jme.2015.2548.
Potočnik, P., Strmčnik, E., & Govekar, E.
(2015).
Linear and Neural Network-based Models for Short-Term Heat Load Forecasting.
Strojniški vestnik - Journal of Mechanical Engineering, 61(9), 543-550.
doi:http://dx.doi.org/10.5545/sv-jme.2015.2548
@article{sv-jmesv-jme.2015.2548,
	author = {Primož  Potočnik and Ervin  Strmčnik and Edvard  Govekar},
	title = {Linear and Neural Network-based Models for Short-Term Heat Load Forecasting},
	journal = {Strojniški vestnik - Journal of Mechanical Engineering},
	volume = {61},
	number = {9},
	year = {2015},
	keywords = {district heating; heat load forecasting; feature extraction; stepwise regression; autoregressive model; neural networks},
	abstract = {Successful operation of a district heating system requires optimal scheduling of heating resources to satisfy heating demands. The optimal operation, therefore, requires accurate short-term forecasts of future heat load. In this paper, short-term forecasting of heat load in a district heating system of Ljubljana is presented. Heat load data and weather-related influential variables for five subsequent winter seasons of district heating operation are applied in this study. Various linear models and nonlinear neural network-based forecasting models are developed to forecast the future daily heat load with the forecasting horizon one day ahead. The models are evaluated based on generalization error, obtained on an independent test data set. Results demonstrate the importance of outdoor temperature as the most important influential variable. Other influential inputs include solar radiation and extracted features denoting population activities (such as day of the week). Comparison of forecasting models reveals good forecasting performance of a linear stepwise regression model (SR) that utilizes only a subset of the most relevant input variables. The operation of the SR model was improved by using neural network (NN) models, and also NN models with a direct linear link (NNLL). The latter showed the overall best forecasting performance, which suggests that NN or the proposed NNLL structures should be considered as forecasting solutions for applied forecasting in district heating markets.},
	issn = {0039-2480},	pages = {543-550},	doi = {10.5545/sv-jme.2015.2548},
	url = {https://www.sv-jme.eu/article/linear-and-neural-network-based-models-for-short-term-heat-load-forecasting/}
}
Potočnik, P.,Strmčnik, E.,Govekar, E.
2015 June 61. Linear and Neural Network-based Models for Short-Term Heat Load Forecasting. Strojniški vestnik - Journal of Mechanical Engineering. [Online] 61:9
%A Potočnik, Primož 
%A Strmčnik, Ervin 
%A Govekar, Edvard 
%D 2015
%T Linear and Neural Network-based Models for Short-Term Heat Load Forecasting
%B 2015
%9 district heating; heat load forecasting; feature extraction; stepwise regression; autoregressive model; neural networks
%! Linear and Neural Network-based Models for Short-Term Heat Load Forecasting
%K district heating; heat load forecasting; feature extraction; stepwise regression; autoregressive model; neural networks
%X Successful operation of a district heating system requires optimal scheduling of heating resources to satisfy heating demands. The optimal operation, therefore, requires accurate short-term forecasts of future heat load. In this paper, short-term forecasting of heat load in a district heating system of Ljubljana is presented. Heat load data and weather-related influential variables for five subsequent winter seasons of district heating operation are applied in this study. Various linear models and nonlinear neural network-based forecasting models are developed to forecast the future daily heat load with the forecasting horizon one day ahead. The models are evaluated based on generalization error, obtained on an independent test data set. Results demonstrate the importance of outdoor temperature as the most important influential variable. Other influential inputs include solar radiation and extracted features denoting population activities (such as day of the week). Comparison of forecasting models reveals good forecasting performance of a linear stepwise regression model (SR) that utilizes only a subset of the most relevant input variables. The operation of the SR model was improved by using neural network (NN) models, and also NN models with a direct linear link (NNLL). The latter showed the overall best forecasting performance, which suggests that NN or the proposed NNLL structures should be considered as forecasting solutions for applied forecasting in district heating markets.
%U https://www.sv-jme.eu/article/linear-and-neural-network-based-models-for-short-term-heat-load-forecasting/
%0 Journal Article
%R 10.5545/sv-jme.2015.2548
%& 543
%P 8
%J Strojniški vestnik - Journal of Mechanical Engineering
%V 61
%N 9
%@ 0039-2480
%8 2018-06-27
%7 2018-06-27
Potočnik, Primož, Ervin  Strmčnik, & Edvard  Govekar.
"Linear and Neural Network-based Models for Short-Term Heat Load Forecasting." Strojniški vestnik - Journal of Mechanical Engineering [Online], 61.9 (2015): 543-550. Web.  01 Jun. 2020
TY  - JOUR
AU  - Potočnik, Primož 
AU  - Strmčnik, Ervin 
AU  - Govekar, Edvard 
PY  - 2015
TI  - Linear and Neural Network-based Models for Short-Term Heat Load Forecasting
JF  - Strojniški vestnik - Journal of Mechanical Engineering
DO  - 10.5545/sv-jme.2015.2548
KW  - district heating; heat load forecasting; feature extraction; stepwise regression; autoregressive model; neural networks
N2  - Successful operation of a district heating system requires optimal scheduling of heating resources to satisfy heating demands. The optimal operation, therefore, requires accurate short-term forecasts of future heat load. In this paper, short-term forecasting of heat load in a district heating system of Ljubljana is presented. Heat load data and weather-related influential variables for five subsequent winter seasons of district heating operation are applied in this study. Various linear models and nonlinear neural network-based forecasting models are developed to forecast the future daily heat load with the forecasting horizon one day ahead. The models are evaluated based on generalization error, obtained on an independent test data set. Results demonstrate the importance of outdoor temperature as the most important influential variable. Other influential inputs include solar radiation and extracted features denoting population activities (such as day of the week). Comparison of forecasting models reveals good forecasting performance of a linear stepwise regression model (SR) that utilizes only a subset of the most relevant input variables. The operation of the SR model was improved by using neural network (NN) models, and also NN models with a direct linear link (NNLL). The latter showed the overall best forecasting performance, which suggests that NN or the proposed NNLL structures should be considered as forecasting solutions for applied forecasting in district heating markets.
UR  - https://www.sv-jme.eu/article/linear-and-neural-network-based-models-for-short-term-heat-load-forecasting/
@article{{sv-jme}{sv-jme.2015.2548},
	author = {Potočnik, P., Strmčnik, E., Govekar, E.},
	title = {Linear and Neural Network-based Models for Short-Term Heat Load Forecasting},
	journal = {Strojniški vestnik - Journal of Mechanical Engineering},
	volume = {61},
	number = {9},
	year = {2015},
	doi = {10.5545/sv-jme.2015.2548},
	url = {https://www.sv-jme.eu/article/linear-and-neural-network-based-models-for-short-term-heat-load-forecasting/}
}
TY  - JOUR
AU  - Potočnik, Primož 
AU  - Strmčnik, Ervin 
AU  - Govekar, Edvard 
PY  - 2018/06/27
TI  - Linear and Neural Network-based Models for Short-Term Heat Load Forecasting
JF  - Strojniški vestnik - Journal of Mechanical Engineering; Vol 61, No 9 (2015): Strojniški vestnik - Journal of Mechanical Engineering
DO  - 10.5545/sv-jme.2015.2548
KW  - district heating, heat load forecasting, feature extraction, stepwise regression, autoregressive model, neural networks
N2  - Successful operation of a district heating system requires optimal scheduling of heating resources to satisfy heating demands. The optimal operation, therefore, requires accurate short-term forecasts of future heat load. In this paper, short-term forecasting of heat load in a district heating system of Ljubljana is presented. Heat load data and weather-related influential variables for five subsequent winter seasons of district heating operation are applied in this study. Various linear models and nonlinear neural network-based forecasting models are developed to forecast the future daily heat load with the forecasting horizon one day ahead. The models are evaluated based on generalization error, obtained on an independent test data set. Results demonstrate the importance of outdoor temperature as the most important influential variable. Other influential inputs include solar radiation and extracted features denoting population activities (such as day of the week). Comparison of forecasting models reveals good forecasting performance of a linear stepwise regression model (SR) that utilizes only a subset of the most relevant input variables. The operation of the SR model was improved by using neural network (NN) models, and also NN models with a direct linear link (NNLL). The latter showed the overall best forecasting performance, which suggests that NN or the proposed NNLL structures should be considered as forecasting solutions for applied forecasting in district heating markets.
UR  - https://www.sv-jme.eu/article/linear-and-neural-network-based-models-for-short-term-heat-load-forecasting/
Potočnik, Primož, Strmčnik, Ervin, AND Govekar, Edvard.
"Linear and Neural Network-based Models for Short-Term Heat Load Forecasting" Strojniški vestnik - Journal of Mechanical Engineering [Online], Volume 61 Number 9 (27 June 2018)

Authors

Affiliations

  • University of Ljubljana, Faculty of Mechanical Engineering, Slovenia 1

Paper's information

Strojniški vestnik - Journal of Mechanical Engineering 61(2015)9, 543-550

https://doi.org/10.5545/sv-jme.2015.2548

Successful operation of a district heating system requires optimal scheduling of heating resources to satisfy heating demands. The optimal operation, therefore, requires accurate short-term forecasts of future heat load. In this paper, short-term forecasting of heat load in a district heating system of Ljubljana is presented. Heat load data and weather-related influential variables for five subsequent winter seasons of district heating operation are applied in this study. Various linear models and nonlinear neural network-based forecasting models are developed to forecast the future daily heat load with the forecasting horizon one day ahead. The models are evaluated based on generalization error, obtained on an independent test data set. Results demonstrate the importance of outdoor temperature as the most important influential variable. Other influential inputs include solar radiation and extracted features denoting population activities (such as day of the week). Comparison of forecasting models reveals good forecasting performance of a linear stepwise regression model (SR) that utilizes only a subset of the most relevant input variables. The operation of the SR model was improved by using neural network (NN) models, and also NN models with a direct linear link (NNLL). The latter showed the overall best forecasting performance, which suggests that NN or the proposed NNLL structures should be considered as forecasting solutions for applied forecasting in district heating markets.

district heating; heat load forecasting; feature extraction; stepwise regression; autoregressive model; neural networks