摘要
飞机在进行结构设计时,需要考虑飞机结构抗炮振设计问题,因此,确定航炮发射时的动载荷大小及时间历程尤为重要。相对于传统的动载荷识别方法,以深度学习技术为支撑的深度神经网络具有强大的拟合能力,对于动载荷的识别具有广阔的应用前景。本文从深度神经网络应用的角度出发,建立航炮炮振载荷识别方法,以与某型炮舱结构动力学相似的简化炮舱模型为研究对象,对复杂波形冲击载荷的动载荷环境进行模拟并对简化炮舱试验模型进行激励;从信号处理的角度将有阻尼动力学系统与有限长脉冲响应系统等效,提取对应的特征信号,应用长短期记忆(LSTM)神经网络对简化炮舱模型试验的冲击动载荷进行识别,并从鲁棒性角度对方法的应用性能进行测试。最终,利用本文所建立的方法对实际的某型炮舱在真实炮振载荷环境下受到的冲击波脉动压力载荷进行了识别,验证了该方法在实际应用场景中的适用性,为炮振动载荷这类复杂冲击载荷的识别提供了新的思路和技术途径。
In the design of aircraft structures,it is necessary to consider the anti-gun vibration design of aircraft structures and to determine the dynamic load when the aircraft machine gun is firing.In comparison with the traditional dynamic load recognition method,deep neural networks supported by deep learning technology have a strong fitting ability and a broad application prospect in dynamic load recognition.This paper presents a vibration load identification method for artillery hands,established from the perspective of deep neural network application.A simplified artillery bay model with similar structural dynamics of a certain type of gun bay was taken as the research object to simulate the dynamic load environment of the complex waveform impact load and the simplified gun bay experimental model.From the perspective of signal processing,the damped dynamic system is equivalent to a finitelength impulse response system.The corresponding feature signals are extracted,and impact dynamic load recognition experiments are carried out on the simplified gun bay model by means of the LSTM neural network.The application performance of the method is examining in terms of robustness.Finally,the method established in this article was used to identify the shock wave load experienced by a certain type of gun compartment in a real gun vibration load environment,verifying the applicability of the method in practical application scenarios and providing new ideas and technical approaches for the identification of complex impact loads such as gun vibration loads.
作者
黄虎
刘翛然
王用岩
杨建
杨智春
Huang Hu;Liu Xiaoran;Wang Yongyan;Yang Jian;Yang Zhichun(National Key Laboratory of Digital and Agile Aircraft Design,AVIC Chengdu Aircraft Design&Research Institute,Chendu 610091,China;National Key Laboratory of Strength and Structural Integrity,Northwestern Polytechnical University,Xi’an 710072,China)
出处
《航空科学技术》
2025年第1期1-10,共10页
Aeronautical Science & Technology
基金
航空科学基金(20220015053002)
关键词
深度学习
载荷识别
振动分析
特征提取
鲁棒性
deep learning
load identification
vibration analysis
feature extraction
robustness
