Chiller power consumption forecasting for commercial building based on hybrid convolution neural networks-long short-term memory model with barnacles mating optimizer
This paper addresses the critical challenge of energy efficiency in commercial buildings, where chillers typically consume 40–50% of total building energy. Accurate forecasting of chiller power consumption is essential for optimizing building energy management systems and reducing operational costs....
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-07-01
|
| Series: | Next Energy |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2949821X25000845 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849387089428217856 |
|---|---|
| author | Mohd Herwan Sulaiman Zuriani Mustaffa |
| author_facet | Mohd Herwan Sulaiman Zuriani Mustaffa |
| author_sort | Mohd Herwan Sulaiman |
| collection | DOAJ |
| description | This paper addresses the critical challenge of energy efficiency in commercial buildings, where chillers typically consume 40–50% of total building energy. Accurate forecasting of chiller power consumption is essential for optimizing building energy management systems and reducing operational costs. Despite advances in deep learning, existing forecasting models often struggle with the complex temporal dependencies and non-linear patterns in chiller operation data. This paper presents an innovative approach using a hybrid Convolutional Neural Network-Long Short-Term Memory (CNN-LSTM) model optimized by the Barnacles Mating Optimizer (BMO). The study compares the proposed CNN-LSTM-BMO against other metaheuristic optimization algorithms, including Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Ant Colony Optimization (ACO), and Differential Evolution (DE). The models were evaluated using comprehensive performance metrics and validated through statistical analysis. Results demonstrate that the CNN-LSTM-BMO achieves superior performance with the lowest Root Mean Square Error (RMSE) of 0.5523 and highest R² value of 0.9435, showing statistically significant improvements over other optimization methods as confirmed by paired t-tests (P < 0.05). Key observations include: (1) the CNN-LSTM-BMO model converges 27% faster than traditional optimization methods; (2) SHapley Additive exPlanations (SHAP) analysis reveals that temperature-related features, particularly saturation temperature, are the most influential predictors across all models; and (3) the proposed model maintains prediction accuracy even under varying operational conditions. The proposed CNN-LSTM-BMO model demonstrates robust convergence characteristics and superior generalization capability, making it particularly suitable for real-world applications in building energy management systems. This research contributes to the advancement of accurate and efficient chiller power consumption forecasting methodologies, offering practical implications for Heating, Ventilation, and Air Conditioning (HVAC) system optimization and energy efficiency improvements in commercial buildings. |
| format | Article |
| id | doaj-art-79c584c2a2b7472c94cfd4639e55baa5 |
| institution | Kabale University |
| issn | 2949-821X |
| language | English |
| publishDate | 2025-07-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Next Energy |
| spelling | doaj-art-79c584c2a2b7472c94cfd4639e55baa52025-08-20T03:55:22ZengElsevierNext Energy2949-821X2025-07-01810032110.1016/j.nxener.2025.100321Chiller power consumption forecasting for commercial building based on hybrid convolution neural networks-long short-term memory model with barnacles mating optimizerMohd Herwan Sulaiman0Zuriani Mustaffa1Faculty of Electrical and Electronics Engineering Technology, Universiti Malaysia Pahang Al-Sultan Abdullah (UMPSA), Pekan, Pahang, 26600, Malaysia; Center for Research in Advanced Fluid and Processes (Fluid Center), Universiti Malaysia Pahang Al-Sultan Abdullah, Gambang Campus, Kuantan, Pahang, 26300, Malaysia; Corresponding author at: Electrical and Electronics Engineering Technology, Universiti Malaysia Pahang Al-Sultan Abdullah (UMPSA), Pekan, Pahang, 26600, Malaysia.Faculty of Computing, Universiti Malaysia Pahang Al-Sultan Abdullah (UMPSA), Pekan, Pahang, 26600, MalaysiaThis paper addresses the critical challenge of energy efficiency in commercial buildings, where chillers typically consume 40–50% of total building energy. Accurate forecasting of chiller power consumption is essential for optimizing building energy management systems and reducing operational costs. Despite advances in deep learning, existing forecasting models often struggle with the complex temporal dependencies and non-linear patterns in chiller operation data. This paper presents an innovative approach using a hybrid Convolutional Neural Network-Long Short-Term Memory (CNN-LSTM) model optimized by the Barnacles Mating Optimizer (BMO). The study compares the proposed CNN-LSTM-BMO against other metaheuristic optimization algorithms, including Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Ant Colony Optimization (ACO), and Differential Evolution (DE). The models were evaluated using comprehensive performance metrics and validated through statistical analysis. Results demonstrate that the CNN-LSTM-BMO achieves superior performance with the lowest Root Mean Square Error (RMSE) of 0.5523 and highest R² value of 0.9435, showing statistically significant improvements over other optimization methods as confirmed by paired t-tests (P < 0.05). Key observations include: (1) the CNN-LSTM-BMO model converges 27% faster than traditional optimization methods; (2) SHapley Additive exPlanations (SHAP) analysis reveals that temperature-related features, particularly saturation temperature, are the most influential predictors across all models; and (3) the proposed model maintains prediction accuracy even under varying operational conditions. The proposed CNN-LSTM-BMO model demonstrates robust convergence characteristics and superior generalization capability, making it particularly suitable for real-world applications in building energy management systems. This research contributes to the advancement of accurate and efficient chiller power consumption forecasting methodologies, offering practical implications for Heating, Ventilation, and Air Conditioning (HVAC) system optimization and energy efficiency improvements in commercial buildings.http://www.sciencedirect.com/science/article/pii/S2949821X25000845Barnacles mating optimizerChiller power consumption forecastingConvolution neural networksLong short-term memoryMetaheuristic algorithms |
| spellingShingle | Mohd Herwan Sulaiman Zuriani Mustaffa Chiller power consumption forecasting for commercial building based on hybrid convolution neural networks-long short-term memory model with barnacles mating optimizer Next Energy Barnacles mating optimizer Chiller power consumption forecasting Convolution neural networks Long short-term memory Metaheuristic algorithms |
| title | Chiller power consumption forecasting for commercial building based on hybrid convolution neural networks-long short-term memory model with barnacles mating optimizer |
| title_full | Chiller power consumption forecasting for commercial building based on hybrid convolution neural networks-long short-term memory model with barnacles mating optimizer |
| title_fullStr | Chiller power consumption forecasting for commercial building based on hybrid convolution neural networks-long short-term memory model with barnacles mating optimizer |
| title_full_unstemmed | Chiller power consumption forecasting for commercial building based on hybrid convolution neural networks-long short-term memory model with barnacles mating optimizer |
| title_short | Chiller power consumption forecasting for commercial building based on hybrid convolution neural networks-long short-term memory model with barnacles mating optimizer |
| title_sort | chiller power consumption forecasting for commercial building based on hybrid convolution neural networks long short term memory model with barnacles mating optimizer |
| topic | Barnacles mating optimizer Chiller power consumption forecasting Convolution neural networks Long short-term memory Metaheuristic algorithms |
| url | http://www.sciencedirect.com/science/article/pii/S2949821X25000845 |
| work_keys_str_mv | AT mohdherwansulaiman chillerpowerconsumptionforecastingforcommercialbuildingbasedonhybridconvolutionneuralnetworkslongshorttermmemorymodelwithbarnaclesmatingoptimizer AT zurianimustaffa chillerpowerconsumptionforecastingforcommercialbuildingbasedonhybridconvolutionneuralnetworkslongshorttermmemorymodelwithbarnaclesmatingoptimizer |