Prediction and analysis of separation performance of on-board hollow fiber membrane
-
摘要:
机载中空纤维膜分离性能实验周期长，操作工况复杂，有必要研究使用较少的数据进行膜性能的精准预测。因此，建立了一种基于人工神经网络的中空纤维膜分离性能预测模型，搭建误差反向传播(BP)神经网络、Elman神经网络和遗传算法优化的BP(GA-BP)神经网络，针对机载中空纤维膜分离性能随引气压力、飞行高度和引气流量的变化规律数据，对比预测不同算法下中空纤维膜的性能预测情况。研究结果表明：搭建的3种神经网络算法均能精准预测中空纤维膜的分离特性，其平均百分比误差(MAPE)均小于1%；分别选取全体数据集的5%、10%和20%，GA-BP神经网络的预测结果最佳，其预测准确率比其他2种神经网络平均高1.43%。基于研究结果，在实际工作中，可以选择合适的算法进行膜实验数据的处理与性能的高效分析。
Abstract:Due to the long experimental period and complex operating conditions, it is necessary to study the precise prediction of membrane performance using limited data in airborne hollow fiber membrane separation performance experiments. An artificial neural network-based prediction model for hollow fiber membrane separation performance has been developed. Error back propagation (BP) neural networks, Elman neural networks, and genetic algorithm optimized BP (GA-BP) neural networks have been constructed to investigate the variation of airborne hollow fiber membrane separation performance with bleed pressure, flight altitude, and bleed flow rate. Compare and predict the performance of hollow fiber membranes under different algorithms. The research results indicate that the three network algorithms constructed can accurately predict the separation characteristics of hollow fiber membranes, with an mean absolute percentage error (MAPE) of less than 1%. With an average prediction accuracy 1.43% greater than the other two, GA-BP neural networks typically produce the best prediction results when selecting 5%, 10%, and 20% of the complete dataset. Based on the research results, suitable algorithms can be selected in practical work to process membrane experimental data and efficiently analyze performance.
-
图 1 BP神经网络模型预测结果与测试集结果比较
Figure 1. Comparison of BP neural network between model prediction results and test set results
图 2 Elman神经网络模型预测结果与测试集结果比较
Figure 2. Comparison of Elman neural network between model prediction results and test set results
图 3 2种神经网络预测值与实际值误差对比
Figure 3. Comparison of error between two neural network predictions and actual values
表 1 隐含层节点数确定结果
Table 1. Result of determining the number of hidden layer nodes
隐含层节点数 MSE 3 $ 2.390\;2\times {10}^{-3} $ 4 $ 4.271\;6\times {10}^{-2} $ 5 $ 2.924\;6\times {10}^{-4} $ 6 $ 7.785\;1\times {10}^{-4} $ 7 $ 2.052\;9\times {10}^{-2} $ 8 $ 9.288\;6\times {10}^{-5} $ 9 $ 1.471\;4\times {10}^{-2} $ 10 $ 8.544\;9\times {10}^{-5} $ 11 $ 1.022\;9\times {10}^{-3} $ 12 $ 1.450\;9\times {10}^{-4} $ 
下载: 导出CSV
表 2 遗传算法参数设置
Table 2. Genetic algorithm parameter settings
种群规模 最大迭代次数 交叉概率 变异概率 各基因取值范围 30 30 0.8 0.2 （−3，3）
下载: 导出CSV
-
[1] 刘卫华, 冯诗愚. 飞机燃油箱惰化技术[M]. 北京: 科学出版社, 2018: 2-3.LIU W H, FENG S Y. Aircraft fuel tank inerting technology[M]. Beijing: Science Press, 2018: 2-3(in Chinese). [2] 贺高红, 邹越, 徐仁贤, 等. 中空纤维气体分离膜丝内压降规律的研究[J]. 膜科学与技术, 1994, 14(2): 22-28.HE G H, ZOU Y, XU R X , et al. A study on pressure drop in hollow-fiber membrane for gas separation[J]. Membrane Science and Technology, 1994, 14(2): 22-28(in Chinese). [3] 冯诗愚, 卢吉, 刘卫华, 等. 机载制氮系统中空纤维膜分离特性[J]. 航空动力学报, 2012, 27(6): 1332-1339.FENG S Y, LU J, LIU W H, et al. Separation performance of hollow fiber membrane for on-board inerting gas generating system[J]. Journal of Aerospace Power, 2012, 27(6): 1332-1339(in Chinese). [4] 蔡琰, 林贵平, 曾宇, 等. 中空纤维膜机载制氮装置的数学建模分析[J]. 航空动力学报, 2015, 30(9): 2100-2107.CAI Y, LIN G P, ZENG Y, et al. Mathematical modeling analysis of hollow fiber membrane onboard inert gas generation system[J]. Journal of Aerospace Power, 2015, 30(9): 2100-2107(in Chinese). [5] YANG J Y, TAI B T, CHAO W C, et al. Analysis of the influence of on-board temperature and pressure control system on inert gas generating performance of hollow fiber membrane[J]. IOP Conference Series: Materials Science and Engineering, 2020, 751(1): 012048. [6] 耿雷铭, 张瑞华, 刘卫华. 中空纤维膜极化系数影响因素分析[J]. 航空动力学报, 2025, 40(3): 14-20.GENG L M, ZHANG R H, LIU W H. Influencing factors of polarization coefficient of hollow fiber membrane[J]. Journal of Aerospace Power, 2025, 40(3): 14-20(in Chinese). [7] 薛勇, 刘卫华, 高秀峰, 等. 机载惰化系统中空纤维膜分离性能的实验研究[J]. 西安交通大学学报, 2011, 45(3): 107-111.XUE Y, LIU W H, GAO X F, et al. Experimental study on separation performance of hollow fiber membrane for onboard inert gas generating system[J]. Journal of Xi’an Jiaotong University, 2011, 45(3): 107-111(in Chinese). [8] 刘小芳, 刘卫华, 钱国诚, 等. 机载中空纤维膜富氮性能实验[J]. 航空动力学报, 2012, 27(5): 975-980.LIU X F, LIU W H, QIAN G C, et al. Experimentation on nitrogen-enriched characteristics of on-board hollow fibre membrane[J]. Journal of Aerospace Power, 2012, 27(5): 975-980(in Chinese). [9] BURNS M, CAVAGE W M, HILL R, et al. Flight-testing of the FAA onboard inert gas generation system on an airbus A320: DOT/FAA/AR-03/58[R]. New Jersey: FAA, 2004. [10] MICHAEL B. Evaluation of fuel tank flammability and the FAA inerting system on the NASA 747 SCA: DOT/FAA/AR-04/41[R]. New Jersey: FAA, 2004. [11] MICHAEL B, WILLIAM M C, RICHARD H, et al. Flight-testing of the FAA onboard inert gas generation system on an airbus A320: DOT/FAA/AR- 03/58[R]. New Jersey: FAA, 2004. [12] 李振环, 王海, 丁小飞, 等. 基于递归径向基神经网络的航空发动机盘腔瞬态壁温预测[J]. 推进技术, 2023, 44(12): 187-195.LI Z H, WANG H, DING X F, et al. Transient wall temperature prediction of aero-engine disk cavity based on recurrent radial basis function neural network[J]. Journal of Propulsion Technology, 2023, 44(12): 187-195. [13] 薛勇. 机载中空纤维膜分离性能及民机燃油箱冲洗惰化研究[D]. 南京: 南京航空航天大学, 2010.XUE Y. Study on separation performance of onboard hollow fiber membrane and flush inerting for fuel tank of civil aircraft[D]. Nanjing: Nanjing University of Aeronautics and Astronautics, 2010(in Chinese). [14] 倪志伟. BP网络中激活函数的深入研究[J]. 安徽大学学报(自然科学版), 1997, 21(3): 48-51.NI Z W. Deep study on activation function in BP network[J]. Journal of Anhui University (Natural Sciences), 1997, 21(3): 48-51(in Chinese). [15] 张学锋, 朱梅, 汤亚玲, 等. 基于GA-BP的烧结风箱阀门预测模型[J]. 西华大学学报(自然科学版), 2025, 44(2): 113-119.ZHANG X F, ZHU M, TANG Y L, et al. GA-BP based sintering airbox valve prediction model[J]. Journal of Xihua University (Natural Science Edition), 2025, 44(2): 113-119(in Chinese). [16] 陈亚琴, 彭浩, 冯诗愚. 基于人工神经网络的微重力流动冷凝换热预测[J]. 南京航空航天大学学报, 2021, 53(6): 989-995.CHEN Y Q, PENG H, FENG S Y. Prediction of flow condensation heat transfer under microgravity flow based on artificial neural network[J]. Journal of Nanjing University of Aeronautics & Astronautics, 2021, 53(6): 989-995(in Chinese). -
下载:
点击查看大图
计量
- 文章访问数: 298
- HTML全文浏览量: 124
- PDF下载量: 0
- 被引次数: 0
