Soil hydraulic properties were mainly governed by soil structures especially when the structures is full of the connected soil macro-pores.Therefore,the good hydrological models need to be well documented for revealin...Soil hydraulic properties were mainly governed by soil structures especially when the structures is full of the connected soil macro-pores.Therefore,the good hydrological models need to be well documented for revealing the process of soil water movement affected by soil medium.The Hydrus-2D model with double domain was recommended in simulating water movement in a heterogeneous medium of soil.To evaluate the performance of the double domain Hydrus-2D model in loess soil,the dynamic of soil wetting front movement in differential loess soil columns under the constant water head were observed and the processes was simulated by Hydrus-2D model under conditions of different soil properties.The results indicated that the Hydrus-2D model was quite good in simulation of loess soil water movements,and the relative errors of simulation results are less than 15%,MRE less than 5%,and R^(2)>0.9.The results provided the appropriate infiltration parameters of loess soil.展开更多
To ensure the safe operation of batteries,accurately obtaining key internal state parameters is essential.However,traditional parameter measurement methods either require opening the battery or long-term measurements,...To ensure the safe operation of batteries,accurately obtaining key internal state parameters is essential.However,traditional parameter measurement methods either require opening the battery or long-term measurements,which are impractical.Therefore,the fixed values are commonly used for these parameters in electrochemical models and have significant limitations.To overcome these limitations,this paper proposes a deep neural network(DNN)based data-driven evaluation method to determine model parameters.By coupling an improved one-dimensional isothermal pseudo-twodimensional(P2D)model with DNN,this study identified concentration-dependent parameters through detailed discharge curve analysis.The results show that the data-driven method can effectively obtain the change trend of concentration-dependent parameters through the charge and discharge curve,and the method can be extended to different battery systems in different discharge rates and aging applications.This work is expected to provide new parameter selection insights for data-driven battery prediction and monitoring models.展开更多
基金This study was funded by the National Natural Science Foundation of China(41471439,41171421)Chinese Academy of Sciences Visiting Professorships for Senior International Scientists(serial number:2013T2Z0027).
文摘Soil hydraulic properties were mainly governed by soil structures especially when the structures is full of the connected soil macro-pores.Therefore,the good hydrological models need to be well documented for revealing the process of soil water movement affected by soil medium.The Hydrus-2D model with double domain was recommended in simulating water movement in a heterogeneous medium of soil.To evaluate the performance of the double domain Hydrus-2D model in loess soil,the dynamic of soil wetting front movement in differential loess soil columns under the constant water head were observed and the processes was simulated by Hydrus-2D model under conditions of different soil properties.The results indicated that the Hydrus-2D model was quite good in simulation of loess soil water movements,and the relative errors of simulation results are less than 15%,MRE less than 5%,and R^(2)>0.9.The results provided the appropriate infiltration parameters of loess soil.
基金supported by National Natural Science Foundation of China(22478239)Science and Technology Commission of Shanghai Municipality(19DZ2271100)National Natural Science Foundation of China(22208208)。
文摘To ensure the safe operation of batteries,accurately obtaining key internal state parameters is essential.However,traditional parameter measurement methods either require opening the battery or long-term measurements,which are impractical.Therefore,the fixed values are commonly used for these parameters in electrochemical models and have significant limitations.To overcome these limitations,this paper proposes a deep neural network(DNN)based data-driven evaluation method to determine model parameters.By coupling an improved one-dimensional isothermal pseudo-twodimensional(P2D)model with DNN,this study identified concentration-dependent parameters through detailed discharge curve analysis.The results show that the data-driven method can effectively obtain the change trend of concentration-dependent parameters through the charge and discharge curve,and the method can be extended to different battery systems in different discharge rates and aging applications.This work is expected to provide new parameter selection insights for data-driven battery prediction and monitoring models.
