超参数优化神经网络驱动的弹炮碰撞动力学建模

Neural Network-based Projectile-barrel Contact/Impact Dynamics Modeling with Hyperparameter Optimization

摘要

针对弹炮异形曲面动态耦合效应引发的接触碰撞建模难题,提出了一种基于数据驱动的身管-弹丸接触碰撞智能建模方法,并开发了相应的模型性能优化方案。鉴于内膛发射极端环境实测困难性,基于有限元仿真模拟手段构建网络模型训练所需样本库,系统涵盖法向初始碰撞速度梯度变化工况;开发超参数混合智能优化框架,通过试错法筛选确定网络激活函数与隐藏层节点分布域,结合遗传算法精调隐藏层节点数与网络初始权值偏置设置;构建全域统计-局部剖析双层评估机制,采用箱线图量化所建数据驱动模型整体预测波动特性,结合泰勒图剖析代表性中位线模型的局部细节特征,系统检验模型预测性能稳定性与泛化能力。仿真结果验证了良好超参数设置在提升模型外插预测性能中的关键作用。所提研究方案不仅可建立性能优异的弹炮复杂表面接触碰撞响应预测模型,还为其他复杂系统的本构模型研究提供了可迁移的方法框架。

Abstract

To address the modeling challenges of contact/impact phenomena caused by the complex contacting surface between barrel and bourrelet, a modeling method for complex surface contact/impact process between barrel and projectile was developed based on artificial neural networks, accompanied by optimization strategies for the enhancement of model performance. Due to the difficulty of conducting practical tests during extreme launching environment, a barrel-projectile contact/impact model was constructed using finite element software. Simulation was employed to obtain the sample dataset required for neural network training under various initial normal contacting velocity conditions. A hybrid optimization framework for hyperparameter tuning was developed, wherein the activation function and the distribution range of hidden layer nodes were determined through trial-and-error screening. The number of hidden layer nodes and the initial weight-bias configuration were fine-tuned using the genetic algorithm. To comprehensively assess the model performance, a dual evaluation mechanism combining global statistical analysis and local detailed assessment was introduced. The overall predictive fluctuation characteristics of the data-driven model were quantified using box plots, while the Taylor diagram was employed to analyze local predictive accuracy of the representative median model, systematically validating the stability and generalization capability of model. Simulation results confirm the critical role of well-optimized hyperparameters in enhancing the extrapolation performance of model. The proposed approach not only establishes an accurate predictive model for barrel-projectile contact/impact responses but also provides a transferable methodological framework for constitutive modeling in other complex systems.

关键词

身管 / 弹丸 / 复杂表面 / 接触碰撞 / 人工神经网络建模

Key words

barrel / projectile / complex surface / contact/impact process / artificial neural network modelling

参考文献

[1] 杨国来, 葛建立, 孙全兆, 等. 火炮振动与控制的发展现状及应用前景[J]. 振动、测试与诊断, 2021, 41(6): 1043-1051.

[2] YANG Guolai, GE Jianli, SUN Quanzhao, et al. Development and application prospects of gun vibration and control[J]. Journal of Vibration, Measurement & Diagnosis, 2021, 41(6): 1043-1051. (in Chinese)

[3] TIAN Q, FLORES P, LANKARANI H M. A comprehensive survey of the analytical, numerical and experimental methodologies for dynamics of multibody mechanical systems with clearance or imperfect joints[J]. Mechanism and Machine Theory, 2018, 122: 1-57.

[4] CHENG Y H, LI Y G, ZHUANG Q Y, et al. Mechanism-informed friction-dynamics coupling GRU neural network for real-time cutting force prediction[J]. Mechanical Systems and Signal Processing, 2024, 221: 111749.

[5] YANG X Y, ZHOU Z X, LI L H, et al. Collaborative robot dynamics with physical human-robot interaction and parameter identification with PINN[J]. Mechanism and Machine Theory, 2023, 189: 105439.

[6] 王锋锋, 杨国来. 迫击炮复合材料座钣多工况多目标优化[J]. 弹道学报, 2024, 36(1): 97-103.

[7] WANG Fengfeng, YANG Guolai. Multi-objective optimization of composite base-plate of mortar under multi-operating conditions[J]. Journal of Ballistics, 2024, 36(1): 97-103. (in Chinese)

[8] 张浩, 李新艳, 黄勇, 等. 导弹飞行过程中动态载荷识别方法[J]. 弹道学报, 2023, 35(1): 8-12.

[9] ZHANG Hao, LI Xinyan, HUANG Yong, et al. Dynamic load identification method during missile flight[J]. Journal of Ballistics, 2023, 35(1): 8-12. (in Chinese)

[10] CHEN L M, XU Z P, LIU S S, et al. Dynamic modeling of a free-piston engine based on combustion parameters prediction[J]. Energy, 2022, 249: 123792.

[11] RASHNO A, FADAEI S. Convolutional neural networks optimization using multi-objective particle swarm optimization algorithm[J]. Information Sciences, 2024, 689: 121443.

[12] WANG J, CHEN B C, YANG W S, et al. BP neural network multi-module green roof thermal performance prediction model optimized based on sparrow search algorithm[J]. Journal of Building Engineering, 2024, 96: 110615.

[13] GAO Y, KASIHMUDDIN M S M, CHEN J, et al. Binary ant colony optimization algorithm in learning random satisfiability logic for discrete hopfield neural network[J]. Applied Soft Computing, 2024, 166: 112192.

[14] 马佳, 陈光宋, 吉磊, 等. 身管内膛与弹丸前定心部接触碰撞响应分析[J]. 弹道学报, 2019, 31(4): 74-81.

[15] MA Jia, CHEN Guangsong, JI Lei, et al. Analysis on contact/impact response between barrel and bourrelet[J]. Journal of Ballistics, 2019, 31(4): 74-81. (in Chinese)

[16] 陈光宋. 弹炮耦合系统动力学及关键参数识别研究[D]. 南京:南京理工大学, 2016.

[17] CHEN Guangsong. The study on the dynamic of the projectile-barrel coupled system and the corresponding key parameters[D]. Nanjing: Nanjing University of Science and Technology, 2016. (in Chinese)

[18] TAYLOR K E. Summarizing multiple aspects of model performance in a single diagram[J]. Journal of Geophysical Research Atmospheres, 2001, 106(D7): 7183-7192.

基金资助

国家自然科学基金(12002065)

国家自然科学基金(12572045)

湖南省自然科学基金(2024JJ5007)

长沙市自然科学基金(kq2402006)

湖南省教育厅科研项目(24B0299)