基于黑鹰优化算法的无链式弹仓参数辨识研究

Research on Parameter Identification of Chainless Magazine Based on Black Eagle Optimizer

摘要

无链式自动弹仓比传统链式供弹具有更快的速率,但其在运动过程中的不确定性程度更大,难以用等效动力学方程准确描述。为了在复杂工况下对某无链式自动弹仓系统的非线性动力学特性有全面的了解,在机理认知和数据驱动协同建模的基础上,建立自动弹仓的不确定动力学模型,利用优化设计思想对模型不确定参数进行辨识。在建模过程中,引入深度神经网络和神经网络微分技术,挖掘系统中各时变参数所隐含的复杂非线性特性;引入Stribeck摩擦模型,对等效摩擦阻力进行建模。该方法既可以保留弹仓动力学模型的传统形式,又充分利用了测试数据所隐含的非线性特征。在辨识过程中采用黑鹰优化算法进行寻优求解,对弹仓在发射中的满载工况下的不确定参数进行辨识,将所得辨识结果与真实结果进行对比。结果表明,辨识所得到的结果与测试结果具有较高的相似度,验证了模型的准确性和辨识方法的有效性,丰富了弹仓的建模理论,在此基础上进行系统分析与控制研究。

Abstract

The chainless automatic magazine is faster than the traditional chain-fed feeding, but it has a higher degree of uncertainty during its motion, making it difficult to accurately describe with an equivalent dynamic equation. In order to gain a comprehensive understanding of the nonlinear dynamic characteristics of a certain chainless magazine system under complex working conditions, an uncertain dynamic model of the magazine system was established based on the collaborative modeling of mechanism cognition and data-driven. The uncertain parameters of the model were identified using the concept of optimized design. In the modeling process, deep neural networks and neural network differentiation techniques were introduced to explore the complex nonlinear characteristics implied by various time-varying parameters in the system; the Stribeck friction model was introduced to model the equivalent friction resistance. This method not only retains the traditional form of the magazine model but also fully utilizes the nonlinear characteristics implied by the test data. During the identification process, the black eagle optimization algorithm was used for optimization and solution seeking, identifying the uncertain parameters of the magazine under full-load conditions during firing, and the results were compared with the actual results. The results show that the identified results have a high degree of similarity with the test results, verifying the accuracy of the model and the effectiveness of the identification method, enriching the modeling theory of the magazine, and laying the foundation for further system analysis and control research.

关键词

无链式弹仓 / 参数辨识 / 深度神经网络 / Stribeck摩擦模型 / 黑鹰优化算法

Key words

chainless magazine / parameter identification / deep neural networks / stribec kfriction model / black eagle optimizer

参考文献

[1] 侯保林, 孙全兆. 武器发射系统自动化技术[M]. 北京: 北京理工大学出版社, 2018.

[2] HOU Baolin, SUN Quanzhao. Automation technology of weapon launching system[M]. Beijing: Beijing University of Technology Press, 2018. (in Chinese)

[3] 陈伟. 一种新型无链供弹系统的动态性能分析[D]. 南京:南京理工大学, 2014.

[4] CHEN Wei. Dynamic performance analysis of a new chainless ammunition supply system[D]. Nanjing: Nanjing University of Science and Technology, 2014. (in Chinese)

[5] 闵杰. 顶置式无链供弹系统结构及动力学分析[D]. 南京:南京理工大学, 2016.

[6] MIN Jie. Structure and dynamic analysis of top-mounted chainless ammunition supply system[D]. Nanjing: Nanjing University of Science and Technology, 2016. (in Chinese)

[7] 文浩, 侯保林, 林瑜斌, 等. 曲线链式回转弹仓动力学模型不确定参数辨识[J]. 兵工学报, 2024, 45(5): 1460-1471.

[8] WEN Hao, HOU Baolin, LIN Yubin, et al. Identification of uncertain parameters in the dynamic model of curved chain rotary magazine[J]. Journal of Military Engineering, 2024, 45(5): 1460-1471. (in Chinese)

[9] 杨宁, 尹强, 李军伟, 等. 某火炮药协调器动力学与控制一体化建模与分析[J]. 弹道学报, 2022, 34(1): 89-97.

[10] YANG Ning, YIN Qiang, LI Junwei, et al. Integrated modelling and analysis of dynamics and control of an artillery charge coordinator[J]. Journal of Ballistics, 2022, 34(1): 89-97. (in Chinese)

[11] 赵洋, 吴茂林. 旋转弹仓的刚体动力学建模及运动特性分析[J]. 弹箭与制导学报, 2014, 34(1): 132-134,202.

[12] ZHAO Yang, WU Maolin. Rigid-body dynamics modelling and kinematic characterization of a rotating magazine[J]. Journal of Bullet and Guidance, 2014, 34(1): 132-134, 202. (in Chinese)

[13] 闫磊. 基于电流检测的自动装填系统故障诊断研究[D]. 太原:中北大学, 2012.

[14] YAN Lei. Research on fault diagnosis of automatic loading system based on current detection[D]. Taiyuan: North Central University, 2012. (in Chinese)

[15] TODOROVSKI L, DZEROSKI S. Integrating knowledge-driven and data-driven approaches to modeling[J]. Ecological Modelling, 2006, 194(1/3): 3-13.

[16] TAMADDON-JAHROMI H R, CHAKSHU N K, SAZONOV I, et al. Data-driven inverse modelling through neural network (deep learning) and computational heat transfer[J]. Computer Methods in Applied Mechanics and Engineering, 2020, 369: 113217.

[17] ZHANG H B, SAN H J, CHEN J P, et al. Black eagle optimizer: a metaheuristic optimization method for solving engineering optimization problems[J]. Cluster Comput, 2024, 27(9): 12361-12393.

基金资助

国家自然科学基金(52405115)