基于神经网络的蓄能器耐久在线智能测试方法Fused Neural Network-Based Online Intelligent Testing Method for Accumulator Durability
陶新洋,周玉婷,孟凯,韩文彬,刘媛,范学
摘要(Abstract)：
为解决传统蓄能器耐久性试验周期长、性能测试效率低的问题，提出一种基于长短期记忆网络（LSTM）与遗传算法优化前馈神经网络（GA-BP）的实时寿命预测融合模型。该模型能够实时监控样件性能状态，并基于MATLAB与LabVIEW平台集成实现智能耐久试验。结果表明，该方法失效识别位置与实测结果一致，全周期性能预测误差小于15%，平均节省约10%的试验时间。
关键词(KeyWords)： 神经网络;蓄能器;实时监测;剩余寿命预测
基金项目(Foundation):
作者(Author): 陶新洋,周玉婷,孟凯,韩文彬,刘媛,范学
DOI: 10.19710/J.cnki.1003-8817.20250219
参考文献(References)：
- [1] GONG W, YU H F, MA H Y, et al. Durability of Concrete with Different Improvement Measures and Its Service Life Prediction in Island and Reef Environment[J]. China Ocean Eng , 2022, 36:947-958.
- [2] HOSSEINZADEH, M, SAMADVAND, H, HOSSEINZADEH, A, et al. Concrete Strength and Durability Prediction through Deep Learning and Artificial Neural Networks[J].Frontiers of Structural and Civil Engineering , 2024, 18:1540-1555.
- [3] ZHOU J, YAO J, HU H H, et al. Accelerated Aging Life Evaluation Method of Silicone Rubber Based on Segmented Nonlinear Arrhenius Model[J]. Materials Research Innovation, 2015, 19:855-860.
- [4]郭琦.基于性能退化数据的可靠性评估方法研究[D].广州:华南理工大学, 2015.
- [5]姚瞬才,李大威.神经网络与深度学习：基于MATLAB的仿真与实现[M].第1版.北京:清华大学出版社,2022.
- [6] CHAI J, LV Z, ZHANG Z, et al. Nonlinear Failure Analysis on Pressurized Cylindrical Shell Using Finite Element Method:Comparison of Theoretical and Experimental Data[J]. Journal of Failure Analysis and Prevention,2022, 22:1000-1010.
- [7]林建忠.流体力学[M].第1版.北京:清华大学出版社,2005.
- [8]赵越.基于遗传蚁群RBF-BP神经网络的LED寿命预测研究[D].重庆:重庆大学, 2021.
- [9] WANG Y L, DU T L, GUO Y Y, et al. A Novel Kalman smoothing(Ks)-Long Short-Term Memory(LSTM)Hybrid Model for Filling in Short-and Long-Term Missing Values in Significant Wave Height[J]. Measurement, 2025,242.