Peak hail impact forces on PV structures using optimized neural networks under wind-hail coupling
Yimi Dai,Taiting Liu,Yixin Li,Ying Xu,Wei Wang
Abstract
Wind-hail disasters often cause severe damage to photovoltaic structures. Accurately predicting the peak
hail impact force under wind-hail conditions is essential for the safe design, structural optimization, and service life
evaluation of photovoltaic systems. In this study, based on the self-developed hail impact simulation integrated device,
extensive wind-hail coupled experiments were conducted to obtain the peak impact force of hail on photovoltaic
structures. Then, a correlation analysis was conducted on the independent and dependent variables. Finally, based on
the machine learning prediction framework proposed in this paper, the BP, PSO-BP, and FA-BP neural network models
were established. The results show that hail velocity exerts the most pronounced effect on the maximum hail impact
force. Conversely, turbulence exhibits an inverse relationship with this peak force. In terms of model robustness and
accuracy, the BP, FA-BP, and PSO-BP models all showcase commendable performance. Notably, the FA-BP model
stands out with the highest robustness and precision, trailed by the PSO-BP model. These findings are expected to offer
a reference for wind-hail resistance experiments on photovoltaic structures and provide insights for predicting the peak
hail impact force under wind-hail coupling conditions.
Key Words
BP neural network; firefly algorithm; photovoltaic structures; PSO algorithm; wind-hail coupled experiments; wind tunnel test
Address
Yimi Dai — School of Civil Engineering, Hunan University of Science and Technology, Xiangtan 411201, Hunan, China
Taiting Liu — School of Civil Engineering, Hunan University of Science and Technology, Xiangtan 411201, Hunan, China
Yixin Li — School of Civil Engineering, Hunan University of Science and Technology, Xiangtan 411201, Hunan, China
Ying Xu — School of Civil Engineering, Hunan University of Science and Technology, Xiangtan 411201, Hunan, China
Wei Wang — School of Civil Engineering, Hunan University of Science and Technology, Xiangtan 411201, Hunan, China
PDF Viewer
Preview is limited to the first 3 pages. Sign in to access the full PDF.
