基于神经网络的航空SOFC-GT混合动力系统功率匹配多目标优化研究  

Multi-objective optimization of power matching for aviation SOFC-GT hybrid power system based on neural network

作　　者：梁爽 毛军逵[1] 贺振宗 梁凤丽 康乐 蒋新勇 LIANG Shuang;MAO Junkui;HE Zhenzong;LIANG Fengi;KANG Le;JIANG Xinyong(College of Energy and Power Engineering,Nanjing University of Aeronautics and Astronautics,Nanjing 210016,China;Collaborative Innovation Center for Advanced Aero-Engine,Beijing 100191,China)

机构地区：[1]南京航空航天大学能源与动力学院,南京210016 [2]先进航空发动机协同创新中心,北京100191

出　　处：《燃气涡轮试验与研究》2025年第2期119-132,共14页Gas Turbine Experiment and Research

基　　金：国家自然科学基金“叶企孙”科学基金(U2341279);江苏省自然科学基金(BK20231445);航空科学基金(20230028052001);航发集团产学研基金(HFZL2022CXY032)。

摘　　要：针对航空固体氧化物燃料电池(SOFC)和涡轮发动机(GT)混合动力系统的功率匹配问题,提出1种基于神经网络的多目标优化方法。通过数值仿真,分析了不同SOFC与GT功率分配对系统性能(系统总质量、系统热效率、系统发电效率、燃油消耗率、CO_(2)排放率等)的影响。采用BP神经网络代理模型,对不同功率匹配方案进行了性能预测,并结合多目标遗传算法与TOPSIS决策方法,优化得出了最佳功率匹配方案(SOFC输出功率为133.092 kW,GT输出功率为66.908 kW)。研究结果表明,优化后的功率匹配方案系统总质量为117.124 kg、燃油消耗率为0.196 kg/s、系统热效率为73.21%。相较于SOFC和GT功率输出相等的方案,优化后的方案系统总质量降低了27.7%。该方法为航空混合动力系统的功率匹配优化设计提供了1种新的思路。Addressing the power matching problem of the solid oxide fuel cell(SOFC)and gas turbine(GT)hybrid power system for aviation,a multi-objective optimization approach based on neural networks was proposed.Numerical simulations were used to analyze the impact of different power distributions between the SOFC and GT on the system performance(including total system weight,thermal efficiency,electrical efficiency,fuel consumption rate,and CO_(2) emission rate).The BP neural network surrogate model was employed to predict the performance of various power matching schemes,and a multi-objective genetic algorithm combined with the TOPSIS decision-making method was used to optimize the power matching solution.The optimal solution was found with the SOFC output power of 133.092 kW and the GT output power of 66.908 kW.The optimization results show that the total system weight is 117.124 kg,the fuel consumption rate is 0.196 kg/s,and the thermal efficiency is 73.21%.Compared to a solution with equal power outputs from the SOFC and GT,the optimized solution reduces total system weight by 27.7%.This study provides a new method for optimizing power matching in aviation hybrid power systems.

关 键 词：固体氧化物燃料电池 涡轮发动机 混合动力 神经网络 功率匹配 多目标优化 

分 类 号：V231.1[航空宇航科学与技术—航空宇航推进理论与工程]