TY - JOUR
T1 - The dynamic hydrogen production yield forecasting model based on the improved discrete grey method
AU - Hu, Yusha
AU - Li, Jigeng
AU - Man, Yi
AU - Ren, Jingzheng
N1 - Funding Information:
This work was supported by Young Scholars Program of Pazhou Lab [No. PLZ2021KF0019 ] and Joint Supervision Scheme with the Chinese Mainland, Taiwan and Macao Universities - Other Chinese Mainland, Taiwan and Macao Universities (Grant No. G-SB3R ).
Publisher Copyright:
© 2022 Hydrogen Energy Publications LLC
PY - 2022/5/15
Y1 - 2022/5/15
N2 - Hydrogen energy is widely regarded as one of the most promising clean energy sources for carbon emissions reduction due to its high-grade energy and complete clean combustion. Since the mechanism of the production process is difficult to elucidate and efficient production cannot be maintained at all times, it has the disadvantages of instability and low efficiency of production processes. The forecasting results can help the biomass based hydrogen production process to adjust the material input in time, thus ensuring a stable and efficient production. Therefore, to solve those problems, the dynamic hydrogen production yield forecasting model based on the Discrete Grey Method (DGM) and the Gradient Boosting Regression Tree (GBRT) has been proposed. The results show that the forecasting results of the GBRT based model have a strong correlation with the input data, while the DGM based model can eliminate some unreasonable forecasting results. Thus, the proposed model can get good forecasting results in different scenarios since its MAE is 0.2. To verify the proposed model, five widely used forecasting models are used as the contrasting models. The forecasting results show that it has the highest precision and has no uncertainty results.
AB - Hydrogen energy is widely regarded as one of the most promising clean energy sources for carbon emissions reduction due to its high-grade energy and complete clean combustion. Since the mechanism of the production process is difficult to elucidate and efficient production cannot be maintained at all times, it has the disadvantages of instability and low efficiency of production processes. The forecasting results can help the biomass based hydrogen production process to adjust the material input in time, thus ensuring a stable and efficient production. Therefore, to solve those problems, the dynamic hydrogen production yield forecasting model based on the Discrete Grey Method (DGM) and the Gradient Boosting Regression Tree (GBRT) has been proposed. The results show that the forecasting results of the GBRT based model have a strong correlation with the input data, while the DGM based model can eliminate some unreasonable forecasting results. Thus, the proposed model can get good forecasting results in different scenarios since its MAE is 0.2. To verify the proposed model, five widely used forecasting models are used as the contrasting models. The forecasting results show that it has the highest precision and has no uncertainty results.
KW - Dynamic forecasting model
KW - Hydrogen production process
KW - Hydrogen production yield forecasting
KW - Improved discrete grey method
KW - Modelling and simulation
UR - https://www.sciencedirect.com/science/article/pii/S0360319922014926
UR - http://www.scopus.com/inward/record.url?scp=85129557568&partnerID=8YFLogxK
U2 - 10.1016/j.ijhydene.2022.04.026
DO - 10.1016/j.ijhydene.2022.04.026
M3 - Journal article
SN - 0360-3199
VL - 47
SP - 18251
EP - 18260
JO - International Journal of Hydrogen Energy
JF - International Journal of Hydrogen Energy
IS - 42
ER -