KPCA-ICOA-BP模型的液体火箭发动机故障诊断  

Liquid Rocket Engine Fault Diagnosis Based on KPCA-ICOA-BP Model

作　　者：孙传鑫 薛薇 许亮[1] SUN Chuanxin;XUE Wei;XU Liang(Tianjin Key Laboratory for Control Theory&Applications in Complicated Systems,School of Electrical Engineering and Automation,Tianjin University of Technology,Tianjin 300384,China;China Academy of Aerospace Science and Innovation,Beijing 100083,China)

机构地区：[1]天津理工大学电气工程与自动化学院,天津市复杂系统控制理论与应用重点实验室,天津300384 [2]中国航天科技创新研究院,北京100083

出　　处：《航天控制》2024年第6期78-84,共7页Aerospace Control

基　　金：国家自然科学基金资助项目(61975151)。

摘　　要：为了提高液体火箭发动机工作的可靠性,针对液体火箭发动机故障诊断问题,提出一种基于核主成分分析(KPCA)和ICOA-BP算法的故障诊断模型。通过KPCA算法对测量参数进行特征提取和降维,保证在特征充足的情况下降低数据的复杂性,减少计算成本,并提出一种改进后的浣熊优化算法(ICOA)优化BP神经网络,旨在提高BP神经网络诊断精度。利用液氧甲烷火箭发动机试车数据对算法进行验证,实验结果表明,ICOA-BP算法相较于COA-BP算法表现出更快的收敛速度和更高的寻优精度。在KPCA特征提取的数据上,ICOA-BP算法诊断准确率可以达到96.5%,相较于BP神经网络和支持向量机(SVM)诊断准确率分别提高3.5%和3%。同粒子群算法(PSO)和遗传算法(GA)相比,ICOA-BP算法展现出更优秀的全局最优解的搜索能力。In order to improve the reliability of the liquid rocket engine,aiming at solving the liquid rocket engine fault diagnosis problem,a kind of fault diagnosis model based on the kernel principal component analysis(KPCA)and the ICOA-BP algorithm is proposed in this paper.The feature extraction and dimensionality degradation of measured parameters are implemented by using KPCA algorithm which ensures sufficient features amount extracted and coming together by reducing the complexities of the data and the computational cost.And an improved Coati optimization algorithm(ICOA)is proposed to optimize the BP neural network,aiming at improving the diagnostic accuracy of the BP neural network.The algorithm is validated by using liquid-oxygen-methane rocket engine test data,and the experimental results show that the ICOA-BP algorithm exhibits faster convergence speed and higher optimization finding accuracy compared to the COA-BP algorithm.The diagnostic accuracy of ICOA-BP algorithm can reach 96.5%on the data extracted from KPCA features,which is respectively 3.5%and 3%higher than the diagnostic accuracy of BP neural network and Support Vector Machine(SVM).Compared with particle swarm algorithm(PSO)and genetic algorithm(GA),ICOA-BP algorithm demonstrates better searching ability for the global optimal solution.

关 键 词：液体火箭发动机 故障诊断 浣熊优化算法 BP神经网络 核主成分分析 

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]