Prognostic for fuel cell based on particle filter and recurrent neural network fusion structure

Xie, Renyou; Ma, Rui; Pu, Sicheng; Xu, Liangcai; Zhao, Dongdong; Huangfu, Yigeng

doi:10.1016/j.egyai.2020.100017

Cited by 70 publications

(20 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Here, the phase change coefficient is 10 times larger than the commonly used in previous simplified studies, 28 which makes possible for the current research to afford. With the development of computer science and artificial intelligence, a better perspective could be expected 33‐36 …”

Section: Methodsmentioning

confidence: 99%

“…With the development of computer science and artificial intelligence, a better perspective could be expected. [33][34][35][36] Specifically, the Courant number, Co = (δtjUj)/ δx < 0.5, was limited to maintain the numerical stability. The Laplacian term, the spatial gradients, velocity field convection, and the passive scalar flux in the equations were discretized with a second-order numerical accuracy, while the temporal evolution was with a one-order numerical accuracy.…”

Section: Computational Domain and Numerical Proceduresmentioning

confidence: 99%

See 1 more Smart Citation

Vapor condensation in reconstructed gas diffusion layers of proton exchange membrane fuel cell

Jiao

2020

Int J Energy Res

View full text Add to dashboard Cite

Summary In the present study, water is considered to be in vapor state when intruding the gas diffusion layers (GDL) of proton exchange membrane fuel cell (PEMFC). A phase change solver was developed in OpenFOAM platform and validated to model the vapor condensation processes in reconstructed GDL. The stochastic method was adopted to reconstruct the 3D GDL while the vapor/condensate water two‐phase flow problem was solved by employing the volume of fluids (VOF) method. Vapor condensation processes in GDL with/without rib were compared under different current densities and ambient temperatures. The temporal varying condensate droplets’ distributions were displayed and the droplets evolution processes were summarized. Vapor condensation processes can be divided into several stages, including small water‐droplets generation, droplets growth, coalescence, accumulation, and condensate water removal. It was found that the condensate water distribution under the rib was entirely different from that under the channel. Water saturation in GDL without rib was lower at the same current density and same ambient temperature. In addition, the vapor condensation was highly sensitive to the local temperature. Growth of the temporal liquid water saturation presented remarkable difference when local temperature varied. The condensate water breakthrough paths were identified. The cyclic release of condensate water droplets into the channel activated the water saturation fluctuations. Water saturation fluctuated wildly in the GDL with rib along the thickness direction, due to the role of rib in impeding the removal of water. Besides, it was noticed that the rib was not beneficial for oxygen diffusion apart from hindering the condensate water transport. A phase change solver was developed in OpenFOAM platform and validated to model the vapor condensation in a realistic structural GDL. The temporal varying condensate droplets’ distributions were displayed and the droplets evolution processes were summarized. The rib was not beneficial for the oxygen diffusion apart from hindering the condensate water transport, which increases the risks of water flooding.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Computational Domain and Numerical Proceduresmentioning

confidence: 99%

Vapor condensation in reconstructed gas diffusion layers of proton exchange membrane fuel cell

Jiao

2020

Int J Energy Res

View full text Add to dashboard Cite

show abstract

“…Recurrent neural networks (RNNs) have been proposed to deal with these time series problems. [226][227][228][229][230][231][232] Fig. 12…”

Section: Predicting the Performance Of Energy Materials In Servicementioning

confidence: 99%

“…12b shows the workflow of a LSTM RNN predicting the temporal evolution of the performance of in-service porous energy materials. Xie et al 228 fused a particle filter (a common mathematical algorithm in signal processing) and the LSTM RNN to predict the lifetime of a PEMFC. This integrated structure showed reasonable prediction when the training phase was large (60% of the dataset).…”

Section: View Article Onlinementioning

confidence: 99%

Towards the digitalisation of porous energy materials: evolution of digital approaches for microstructural design

Niu

Pinfield

et al. 2021

Energy Environ. Sci.

View full text Add to dashboard Cite

show abstract

“…The operation conditions including fuel cell temperature, relative humidity, hydrogen pressure, and current are considered in the method. In recent years, long shortterm memory recurrent network (LSTM) has also been successfully applied in the performance prediction of PEMFC [22][23][24] . The results demonstrate that LSTM model can achieve the prediction of PEMFC performance with high accuracy.…”

Section: Introductionmentioning

confidence: 99%

Performance prediction of fuel cells using long short‐term memory recurrent neural network

Hou

Tao

et al. 2021

Int J Energy Res

View full text Add to dashboard Cite

Performance prediction of proton-exchange membrane fuel cell (PEMFC) under dynamic conditions, especially for vehicle applications, is increasingly become the focus of attention. This article proposes a performance prediction method of PEMFC using long short-term memory (LSTM) recurrent neural network (RNN). In this article, polarization curve (current-voltage curve) and voltage degradation curve (current-time curve) are adopted as the main performance indexes of PEMFC. Both polarization curve prediction and performance degradation prediction of PEMFC can be effectively implemented based on the LSTM method. To investigate the voltage losses law of experimental and predicted results, the paper introduces an empirical equation of polarization curve. The perfect match between the experimental and predicted polarization losses of PEMFC can further validate the prediction performance of LSTM method. The proposed prediction method is validated by the PEMFC polarization curve data obtained from the designed aging experiment of a 4 kW stack operated under dynamic loading cycling situation during 600 hours. Then, LSTM network is compared with traditional RNN and back-propagation neural network (BPNN) to prove its superiority. The minimum values of rootmean-square error (RMSE) and the mean absolute percentage error (MAPE) of LSTM network with different training data are 0.0088 and 0.0101, respectively. All the coefficient of determination (R 2 ) of LSTM model with different training data is over 0.95, which is close to 1.0. The prediction accuracy of LSTM network is higher than that of two other networks. The result indicates that LSTM network outperforms two other networks in PEMFC performance prediction.Hence, the prediction method based on LSTM network is very suitable for PEMFC performance prediction.

show abstract

Prognostic for fuel cell based on particle filter and recurrent neural network fusion structure

Cited by 70 publications

References 20 publications

Vapor condensation in reconstructed gas diffusion layers of proton exchange membrane fuel cell

Vapor condensation in reconstructed gas diffusion layers of proton exchange membrane fuel cell

Towards the digitalisation of porous energy materials: evolution of digital approaches for microstructural design

Performance prediction of fuel cells using long short‐term memory recurrent neural network

Contact Info

Product

Resources

About