王虎 尹澤泉 王雯婕 黃笠舟 方寧寧 隋俊友 張加樂 張銳明 隋邦傑



摘要:針對質子交換膜燃料電池氣體擴散層(gas diffusion layer composition,GDL)形貌劃分與制備工藝改進問題,提出了一種基于金字塔池化網絡(pyramid scene parsing network,PSPNet)與多層感知器(multi-layer perception,MLP)的氣體擴散層組分識別與比例推理方法:首先將帶標簽的氣體擴散層掃描電鏡(scanning electron microscope,SEM)圖片輸入神經網絡,得到特征圖;得到的圖像特征層進入金字塔池化模塊后,獲取SEM圖像的深層和淺層特征;隨后將深層和淺層特征圖層融合輸入全卷積網絡(fully convolutional network,FCN)模塊,得到預測圖像;最后統計各個組分上的像素點比例,通過MLP完成組分比例推理。結果表明:所提方法組分識別像素準確率達81.24%;在5%偏差范圍內,比例推理準確率為88.89%。該方法解決了氣體擴散層多組分無法區分、比例無法獲知的問題,可有效應用于氣體擴散層的質檢、數值重構以及制備工藝改進。
關鍵詞:質子交換膜燃料電池;氣體擴散層制備;掃描電鏡;人工智能;金字塔池化網絡;多層感知器
中圖分類號:TK91????????? 文獻標志碼:A?????????? 文章編號:1000-582X(2024)01-084-09
Inference method of proton exchange membrane fuel cell gas diffusion layer composition based on pyramid scene parsing network
WANG Hu1,2, YIN Zequan1,3, WANG Wenjie2, HUANG Lizhou3, FANG Ningning3,
SUI Junyou2, ZHANG Jiale4, ZHANG Ruiming5,6, SUI PangChieh1
(1. School of Automotive Engineering, Wuhan University of Technology, Wuhan 430070, P. R. China; 2. Nanjing Royali Intelligent Technology Co.,Ltd., Nanjing 211106, P. R. China; 3. China Automotive Innovation Corporation, Nanjing 211100, P. R. China; 4. College of Aerospace Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, P. R. China; 5. Guangdong Hydrogen Energy Institute of WHUT, Foshan, Guangdong 528200, P. R. China; 6. Foshan Xianhu Laboratory of the Advanced Energy Science and Technology Guangdong Laboratory, Foshan, Guangdong 528200, P. R. China)
Abstract: To rapidly determine the morphology of the gas diffusion layer for proton exchange membrane fuel cell (PEMFC) and improve its fabrication process, a method of gas diffusion layer (GDL) component identification and proportional reasoning based on a combination of pyramid scene parsing network (PSPNet) and multilayer perceptron (MLP) is proposed. First, labeled GDL scanning electron microscope (SEM) images are input into the neural network to obtain a feature extraction map. This map is used in the pyramid pooling module to extract both deep and shallow features of the SEM images. Subsequently, these feature layers are input into the fully convolutional network (FCN) module to produce a predicted image of the same size. Finally, the proportion of pixels for each component is calculated, and the inference of component proportion is achieved by using the MLP. The accuracy of the proposed method is 81.24%, with an accuracy of proportional reasoning reaching 88.89% within a 5% deviation range. The proposed method can be effectively used for gas diffusion layer quality detection, numerical reconstruction, and process improvement.
Keywords: proton exchange membrane fuel cell; gas diffusion layer preparation; scanning electron microscope; artificial intelligence; pyramid scene parsing network; multilayer perceptron
氣體擴散層 (gas diffusion layer,GDL)作為質子交換膜燃料電池最重要的零部件之一,主要負責傳輸反應氣體[1]、導出電子[2]、排出反應生成的水[3-4]。GDL由碳纖維、粘接劑、疏水劑通過高溫石墨化制備而成,各組分的比例及分布情況決定了氣體擴散層的孔隙結構、連通性,從而影響導氣、導電與傳熱等過程,并進一步影響整個膜電極乃至燃料電池的性能。因此,實現對氣體擴散層微觀結構中不同組分的精準識別具有重要意義。……