重庆理工大学学报（自然科学版）2025，Vol.39Issue(3)：227-233,7.DOI:10.3969/j.issn.1674-8425(z).2025.02.028
面向PEMFC长期老化预测的多尺度卷积神经网络
Multi-scale convolutional neural network for long-term aging prediction of PEMFC
摘要
Abstract
Accurate prediction of the long-term aging trends of proton exchange membrane fuel cells(PEMFC)is crucial for system control and diagnostics.However,most data-driven models primarily focus on short-term predictions,making it challenging to deliver precise long-term results.In this paper,we propose a multi-scale convolutional neural network(MCNN)model that leverages multi-scale decomposition to separate the input data into linear and nonlinear trends.This approach enables the model to accurately extract long-term dependencies and effectively capture nonlinear information within the data.By integrating these decomposed trends,the model enhances the accuracy of long-term predictions.The effectiveness and accuracy of our proposed model is validated by the IEEE PHM 2014 fuel cell durability test data.关键词
质子交换膜燃料电池/长期老化预测/数据驱动/多尺度卷积Key words
proton exchange membrane fuel cells/long-term aging prediction/data-driven/multi-scale convolution分类
动力与电气工程引用本文复制引用
谢永平,胡成玉..面向PEMFC长期老化预测的多尺度卷积神经网络[J].重庆理工大学学报（自然科学版）,2025,39(3):227-233,7.基金项目
国家自然科学基金面上项目(62073300) （62073300）
湖北省教育科学规划重点课题(2022GA135) （2022GA135）
