Research on an hourly heat load forecasting model for district heating systems based on heterogeneous model fusion

Accurate load forecasting for district heating systems (DHS) is crucial for ensuring efficient production, distribution, and energy utilization. Although artificial neural networks have been widely used for DHS thermal load forecasting, their capacity to capture long-term dependencies remains limite...

Full description

Saved in:
Bibliographic Details
Main Authors: Tieliu Jiang, Wenyue Liu, Jianqing Lin, Xu Jin, Zhongyan Liu
Format: Article
Language:English
Published: Elsevier 2025-09-01
Series:Case Studies in Thermal Engineering
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2214157X25007609
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849417478794379264
author Tieliu Jiang
Wenyue Liu
Jianqing Lin
Xu Jin
Zhongyan Liu
author_facet Tieliu Jiang
Wenyue Liu
Jianqing Lin
Xu Jin
Zhongyan Liu
author_sort Tieliu Jiang
collection DOAJ
description Accurate load forecasting for district heating systems (DHS) is crucial for ensuring efficient production, distribution, and energy utilization. Although artificial neural networks have been widely used for DHS thermal load forecasting, their capacity to capture long-term dependencies remains limited. This study proposes a prediction model called TBMK (TCN-BiMamba2-KAN) based on an improved bidirectional selective state-space model (Bi-Mamba2). By integrating the local feature extraction capabilities of temporal convolutional networks (TCN), the global temporal modeling advantages of Bi-Mamba2, and the nonlinear fitting characteristics of Kolmogorov-Arnold networks (KAN), a multi-module collaborative prediction framework is constructed. Unlike conventional SSM-based models that struggle with long-term thermal inertia modeling, the proposed TBMK achieves dual-dimensional feature capture through bidirectional state-space scanning and spline-activated nonlinear mapping. The model is validated using historical operational data from a DHS in Changchun, China, which includes multidimensional features such as secondary heating/return water temperatures, solar radiation, humidity, and outdoor temperature. A comparison was conducted with benchmark models, including LSTM, TCN, Bi-Mamba2, TCN-LSTM, and BiMamba2-TCN. The experimental results indicate that the R2 value of the TBMK model reaches 0.987, with the MAE, RMSE, and MAPE decreasing by 25 %, 30.49 %, and 1.17 %, respectively, compared to the second-best model. Diebold-Mariano test confirms that its predictive performance is significantly superior to the comparative models (p < 0.05). This model can provide high-precision guidance for thermal scheduling, facilitating optimization of energy losses and control of operational costs.
format Article
id doaj-art-57e2b940b0bb496f9510f6e63c2b871a
institution Kabale University
issn 2214-157X
language English
publishDate 2025-09-01
publisher Elsevier
record_format Article
series Case Studies in Thermal Engineering
spelling doaj-art-57e2b940b0bb496f9510f6e63c2b871a2025-08-20T03:32:50ZengElsevierCase Studies in Thermal Engineering2214-157X2025-09-017310650010.1016/j.csite.2025.106500Research on an hourly heat load forecasting model for district heating systems based on heterogeneous model fusionTieliu Jiang0Wenyue Liu1Jianqing Lin2Xu Jin3Zhongyan Liu4School of Energy and Power Engineering, Northeast Electric Power University, Jilin, 132012, ChinaSchool of Energy and Power Engineering, Northeast Electric Power University, Jilin, 132012, ChinaSchool of Energy and Power Engineering, Northeast Electric Power University, Jilin, 132012, ChinaSchool of Energy and Power Engineering, Northeast Electric Power University, Jilin, 132012, ChinaCorresponding author.; School of Energy and Power Engineering, Northeast Electric Power University, Jilin, 132012, ChinaAccurate load forecasting for district heating systems (DHS) is crucial for ensuring efficient production, distribution, and energy utilization. Although artificial neural networks have been widely used for DHS thermal load forecasting, their capacity to capture long-term dependencies remains limited. This study proposes a prediction model called TBMK (TCN-BiMamba2-KAN) based on an improved bidirectional selective state-space model (Bi-Mamba2). By integrating the local feature extraction capabilities of temporal convolutional networks (TCN), the global temporal modeling advantages of Bi-Mamba2, and the nonlinear fitting characteristics of Kolmogorov-Arnold networks (KAN), a multi-module collaborative prediction framework is constructed. Unlike conventional SSM-based models that struggle with long-term thermal inertia modeling, the proposed TBMK achieves dual-dimensional feature capture through bidirectional state-space scanning and spline-activated nonlinear mapping. The model is validated using historical operational data from a DHS in Changchun, China, which includes multidimensional features such as secondary heating/return water temperatures, solar radiation, humidity, and outdoor temperature. A comparison was conducted with benchmark models, including LSTM, TCN, Bi-Mamba2, TCN-LSTM, and BiMamba2-TCN. The experimental results indicate that the R2 value of the TBMK model reaches 0.987, with the MAE, RMSE, and MAPE decreasing by 25 %, 30.49 %, and 1.17 %, respectively, compared to the second-best model. Diebold-Mariano test confirms that its predictive performance is significantly superior to the comparative models (p < 0.05). This model can provide high-precision guidance for thermal scheduling, facilitating optimization of energy losses and control of operational costs.http://www.sciencedirect.com/science/article/pii/S2214157X25007609District heating systemsThermal load forecastingBi-Mamba2State space modelTemporal convolutional networkKolmogorov-arnold network
spellingShingle Tieliu Jiang
Wenyue Liu
Jianqing Lin
Xu Jin
Zhongyan Liu
Research on an hourly heat load forecasting model for district heating systems based on heterogeneous model fusion
Case Studies in Thermal Engineering
District heating systems
Thermal load forecasting
Bi-Mamba2
State space model
Temporal convolutional network
Kolmogorov-arnold network
title Research on an hourly heat load forecasting model for district heating systems based on heterogeneous model fusion
title_full Research on an hourly heat load forecasting model for district heating systems based on heterogeneous model fusion
title_fullStr Research on an hourly heat load forecasting model for district heating systems based on heterogeneous model fusion
title_full_unstemmed Research on an hourly heat load forecasting model for district heating systems based on heterogeneous model fusion
title_short Research on an hourly heat load forecasting model for district heating systems based on heterogeneous model fusion
title_sort research on an hourly heat load forecasting model for district heating systems based on heterogeneous model fusion
topic District heating systems
Thermal load forecasting
Bi-Mamba2
State space model
Temporal convolutional network
Kolmogorov-arnold network
url http://www.sciencedirect.com/science/article/pii/S2214157X25007609
work_keys_str_mv AT tieliujiang researchonanhourlyheatloadforecastingmodelfordistrictheatingsystemsbasedonheterogeneousmodelfusion
AT wenyueliu researchonanhourlyheatloadforecastingmodelfordistrictheatingsystemsbasedonheterogeneousmodelfusion
AT jianqinglin researchonanhourlyheatloadforecastingmodelfordistrictheatingsystemsbasedonheterogeneousmodelfusion
AT xujin researchonanhourlyheatloadforecastingmodelfordistrictheatingsystemsbasedonheterogeneousmodelfusion
AT zhongyanliu researchonanhourlyheatloadforecastingmodelfordistrictheatingsystemsbasedonheterogeneousmodelfusion