Dynamic carbon emissions optimization method for HIES based on cloud-edge collaborative CBAM-BiLSTM-PSO network DOI

Songqing Cheng,

Tong Nie, Qian Hui

et al.

Published: May 2, 2025

Abstract To achieve the low carbon optimization in hydrogen-based integrated energy system(HIES), this paper proposes a dynamic emissions method for HIES based on cloud-edge collaborative CBAM-BiLSTM-PSO network. Firstly, theory of emission flow, are converted from source to multiple load nodes, and reduction model is established. The coordinated achieved by setting edge objective function at cloud function. And noise sources correlate relationship between input variables decision variables, uncertainty embedding achieved. Then, computing network established prediction new power output multi-energy consuming as well scheduling plan solving. Convolutional block attention module (CBAM) used strengthen key feature data fuse heterogeneous data. particle swarm algorithm (PSO) combined with bidirectional long short-term memory (BiLSTM) form solving algorithm, which realizes solution plan. Finally, proposed was validated using actual running an example. results showed that can effectively extract operating characteristics equipment within HIES, reduction, reduce HIES. Compared other models, training time shortened accuracy improved, providing feasible data-based low-carbon operation

Language: Английский

Dynamic carbon emissions optimization method for HIES based on cloud-edge collaborative CBAM-BiLSTM-PSO network DOI

Songqing Cheng,

Tong Nie, Qian Hui

et al.

Published: May 2, 2025

Abstract To achieve the low carbon optimization in hydrogen-based integrated energy system(HIES), this paper proposes a dynamic emissions method for HIES based on cloud-edge collaborative CBAM-BiLSTM-PSO network. Firstly, theory of emission flow, are converted from source to multiple load nodes, and reduction model is established. The coordinated achieved by setting edge objective function at cloud function. And noise sources correlate relationship between input variables decision variables, uncertainty embedding achieved. Then, computing network established prediction new power output multi-energy consuming as well scheduling plan solving. Convolutional block attention module (CBAM) used strengthen key feature data fuse heterogeneous data. particle swarm algorithm (PSO) combined with bidirectional long short-term memory (BiLSTM) form solving algorithm, which realizes solution plan. Finally, proposed was validated using actual running an example. results showed that can effectively extract operating characteristics equipment within HIES, reduction, reduce HIES. Compared other models, training time shortened accuracy improved, providing feasible data-based low-carbon operation

Language: Английский

Citations

0