Data-driven prediction of dimensionless quantities for semi-infinite target penetration by integrating machine-learning and feature selection methods

This study employs a data-driven methodology that embeds the principle of dimensional invariance into an artificial neural network to automatically identify dominant dimensionless quantities in the penetration of rod projectiles into semi-infinite metal targets from experimental measurements.The derived mathematical expressions of dimensionless quantities are simplified by the examination of the exponent matrix and coupling relationships between feature variables.As a physics-based dimension reduction methodology,this way reduces high-dimensional parameter spaces to descriptions involving only a few physically interpretable dimensionless quantities in penetrating cases.Then the relative importance of various dimensionless feature variables on the penetration efficiencies for four impacting conditions is evaluated through feature selection engineering.The results indicate that the selected critical dimensionless feature variables by this synergistic method,without referring to the complex theoretical equations and aiding in the detailed knowledge of penetration mechanics,are in accordance with those reported in the reference.Lastly,the determined dimensionless quantities can be efficiently applied to conduct semi-empirical analysis for the specific penetrating case,and the reliability of regression functions is validated.

Investigation on the penetration of jacketed rods with striking velocities of 0.9-3.3 km/s into semi-infinite targets

In this study, a combined experimental, numerical and theoretical investigation is conducted on the penetration of semi-infinite 4340 steel targets by a homogeneous 93 W rod and two types of jacketed rods with striking velocities of 0.9-3.3 km/s. The results show that the jacketed rods produced typical“co-erosion” damage at all test velocities, except for the 93 W/1060 Al jacketed rod, which switched from an early “bi-erosion” damage to later “co-erosion” damage at a striking velocity of 936 m/s. However, the homogeneous 93 W rod always forms a large mushroom head during the penetration process. The damage mechanisms of these two types of jacketed rods differ for striking velocities of 0.9-2.0 km/s, but this difference gradually decreases with increased striking velocity. For velocities of 2.0-3.3 km/s, all three types of projectiles exhibit typical hydrodynamic penetration characteristics, and the damage mechanisms of the two types of jacketed rods are almost identical. For the same initial kinetic energy, the penetration performance of the jacketed rods is distinctly superior to that of the homogeneous 93 W rods.Compared with jacket density, jacket strength shows a more significant influence on the damage mechanism and penetration performance of the jacketed rod. Finally, an existing theoretical prediction model of the penetration depth of jacketed rods on semi-infinite targets in the co-erosion mode is modified. It transpires that-in terms of penetration depth-the modified theoretical model is in good agreement with the experimental and numerical observations for 93 W/TC4 and 93 W/1060 Al jacketed rods penetrating semi-infinite 4340 steel targets.

Analytical models for the penetration of semi-infinite targets by rigid,deformable and erosive long rods

A theoretical study is presented herein on the pen- etration of a semi-infinite target by a spherical-headed long rod for Yp 〉 S, where Yp is the penetrator strength and S is the static target resistance. For Yp 〉 S, depending upon initial impact velocity, there exist three types of penetration, namely, penetration by a rigid long rod, penetration by a deforming non-erosive long rod and penetration by an erosive long rod. If the impact velocity of the penetrator is higher than the hydrodynamic velocity （VH）, it will penetrate the target in an erosive mode; if the impact velocity lies between the hydrodynamic velocity （VH） and the rigid body velocity （VR）, it will penetrate the target in a deformable mode; if the impact velocity is less than the rigid body velocity （VR）, it will penetrate the target in a rigid mode. The critical conditions for the transition among these three penetration modes are proposed. It is demonstrated that the present model predictions correlate well with the experimental observations in terms of depth of penetration （DOP） and the critical transition conditions.

Experimental and numerical study on the influence of shaped charge liner cavity filing on jet penetration characteristics in steel targets

Penetration characteristic(size and shape of penetration craters made in high hardness ARMSTAL 30PM steel) of shaped charge jets formed after detonations of modified PG-7VM warheads was analyzed in the article. Modifications consisted in removing the frontal part of the grenade(fuse, ballistic cap and conductive cone) and introducing of the liner cavity filling made of polyacetal copolymer POM-C. The filings in the form of solid cones with three different heights(33%, 66% and 100% of H-the height of original PG-7VM liner) were placed inside of the hollow cone shaped charge liner. As opposed to the vast majority of previously published works(in which warhead optimization studies were focused on increasing of the depth of penetration in rolled homogeneous armor steel) the main aim of the presented modifications was to maximize the damage ratio(diameters of craters, inlet and outlet holes) of target perforated by shaped charge jet at the cost of the loss of part of the jet penetration capability. According to the best knowledge of the authors such approach to the use of the old PG-7VM warheads has not been analyzed so far. Taking into consideration high stock levels of PG-7VM warheads, and the fact that they are continuously being replaced by more efficient and more sophisticated high-explosive anti-tank warheads, it seems reasonable to look for alternate applications of the warheads withdrawn from the service. Thanks to the introduction of proposed modifications the warheads could be used by special forces or other assault units as directional mines or statically detonated cutting shaped charges as well as by combat engineers as universal charges used in various types of engineering or sapper works. The research included experimental penetration tests and their numerical reproduction in the LS-Dyna software with the simulation methodology defined and validated in previous works of the authors.Small differences(average error = 10-20%) were identified between the experimental and numerical results(dimensions of craters made in steel targets were compared) what confirmed the reliability of the modelling methodology and enabled its use for further optimization of the shapes of fillings. Within the analyzed variants of warheads modifications maximum diameters of penetration craters were obtained for the filling of the height of h = 2/3H. The diameters of holes in individual steel plates were increased by 164%, 70%, 65%(for the first, second and third plate, respectively) in relation to the variant without filling. The results of the study indicated that with the use of different materials of fillings and their various heights it is possible to control the shape of penetration craters pierced in the steel targets.

Selection and modification of ground motion records using Newmark-Hall spectrum as target spectrum for long-period structures

Input ground motions have significant impacts on the uncertainty of structural responses in time-history analysis.In this study,records were selected and scaled for the evaluation of mean structural responses according to the target spectrum.The Newmark-Hall spectrum is closely related to seismic response of short,medium and long-period structures,so it was taken as the target spectrum here.The nonlinear time-history analyses of 9-story and 20-story steel moment-resisting frame structures were carried out as examples.They represent medium and long-period buildings,respectively.Three target spectra with risk of 50%,10%and 2%probabilities for exceedance in 50 years were calculated by the average Newmark-Hall spectrum method for three ground motion sets developed in the SAC Steel Project.The predicted structural mean responses of these Newmark-Hall spectra were compared with those calculated by the average spectral acceleration method for the same record set.It is found that both methods have similar accuracy for estimating the structural mean response.However,the method proposed herein is more effective in reducing the variability of the structural responses.Also,the proposed method is more advantageous for the time-history analysis of long-period structures or structures with more severe nonlinear responses under strong seismic excitations.

Iron and Steel:Hitting the 90-Million-Ton Target

The story of iron and steel production in China is one of constant and spectacular rise,but no one seemns to know yet if the story will hawe a happy ending.

两种钛合金半穿甲战斗部侵彻单层钢靶性能试验研究

选取亚稳态β型TB17和高损伤容限α+β型TC4-DT两种高韧钛合金材料,对2种钛合金材料在高应变率条件下动态力学响应特性进行测试分析,并应用125 mm火炮发射装置开展2种钛合金半穿甲战斗部侵彻单层钢靶性能试验研究。结果表明,作为半穿甲战斗部的壳体材料,应选择屈服强度与冲击韧性匹配性好的钛合金,其中更应关注钛合金材料的塑性变形能力和抗冲击韧性;随着侵彻着角的增加,侵彻单层钢靶过程弹体头部侵蚀量及径向墩粗程度更严重,不利于弹体保持侵彻强度及装药安定性。

加强筋位置对半穿甲战斗部侵彻多层钢靶性能影响研究

为了研究加强筋位置对半穿甲战斗部侵彻多层钢靶性能的影响,通过试验及数值仿真,研究了加强筋在弹体轴向不同位置处的2种半穿甲战斗部,以一定速度(710 m/s)、着角(20°)侵彻8层钢靶的弹道偏转、弹孔尺寸、余速的差异。研究结果表明:加强筋位置由弹尖向质心位置逼近时弹道偏转角减小、弹孔长轴减小、穿甲能力提升;弹道偏转角增量与靶间距呈正相关关系;弹道偏转角、侵彻余速、弹头侵蚀量及弹孔仿真结果与试验实测值最大偏差为8.82%,仿真与试验结果符合较好。研究结果可为半穿甲战斗部方案设计提供参考。

基于改进YOLOv5s 的药盒钢印日期识别方法

目的药盒的钢印日期与背景对比度低,字符轮廓不明显,识别易受环境光线干扰,对此提出一种基于机器视觉的识别方法。方法使用改进YOLOv5s模型,首先对采集的药盒数据集进行透视变换校正,并进行数据增强。通过在模型的骨干网络中融合位置注意力机制(CA),减少冗余信息的干扰;颈部网络根据加权双向特征金字塔网络(BiFPN)引入权重,更好地平衡不同尺寸图层的特征信息;引入动态聚焦损失函数(WIoU),降低高质量样本对训练的干预,提高模型的泛化能力。结果在自建钢印字符数据集上的实验结果表明,改进网络对药盒钢印日期识别的平均精度值达到了99.41%,比原始模型提升了2.38%,帧率为80.01帧/s。结论改进后的YOLOv5模型对药盒钢印日期的检测精度优于原有网络,对可以满足药盒生产线的实时性要求。

高钛高强钢钛含量精准控制研究

对高钛高强钢的冶炼流程进行了分析,针对存在的问题进行了理论与实践相结合的研究,制定相应的改进措施。通过控制真空室内的氧和钢液中的氧含量,提高高钛钢铸机成品目标值w(Ti)±0.005%合格率。

基于机器识别的带钢表面缺陷检测研究

针对传统带钢表面缺陷检测技术落后、效率不高及小目标识别能力不足等问题,提出一种改进的YOLOv5s-Tiny目标检测模型,在保持模型较小计算量的同时提升检测速度和识别精度。通过将主干网络GSP-Darknet53替换为轻量级GhostNet网络,减少模型参数的数量,提高推理速度。在主干网络加入CBAM注意力机制,通过通道注意力机制和空间注意力机制对特征信息进行融合增强,提高小目标检测精度,并将损失函数GIoU改进为EIoU,提高检测框定位能力。最后将改善后的训练模型格式转换后安装到手机安卓端验证优化的有效性。结果表明:在东北大学数据集中,改进后模型检测精度提高1.5%的同时,召回率提升了1.5%,参数量减少12.3%;安卓端检测速度约为120 ms,完成带钢缺陷的实时检测。

钢材表面缺陷检测研究综述

钢材是工业领域不可或缺的原材料,表面缺陷严重影响钢材质量。传统钢材表面缺陷检测方法精度低、速度慢、劳动强度大,无法满足实际生产需求。近年来深度学习技术发展迅速,其能充分挖掘目标图像底层特征信息,给钢材缺陷检测带来了新的解决方案。综述近年钢材表面缺陷检测方法相关文献,简述传统检测方法的原理及其适用性,分析深度学习检测模型的结构与特点,并对目前该领域存在的一些技术难点进行总结,对未来发展趋势进行展望。

基于数据深度增强的钢桥螺栓脱落智能检测方法研究

针对钢桥螺栓脱落病害人工巡检效率低和智能化检测样本数据集不足的问题,提出一种基于数据深度增强的钢桥螺栓脱落智能检测方法。该方法首先以采集的螺栓图像数据集为基础,利用深度卷积生成对抗网络(Deep Convolutional Generative Adversarial Networks,DCGAN)对有限的螺栓图像进行增广;然后将生成的图像与原始图像合并构建增广后的数据集;再将数据集输入到单阶段目标检测网络YOLO(You Only Look Once,YOLO)中,结合迁移学习方法进行模型训练,并对训练后模型的性能进行验证;最后,进行螺栓脱落病害识别。为验证该方法的可行性,对螺栓脱落检测模型进行试验验证,并对不同采集环境下的某钢桥拼接板螺栓脱落病害进行检测。结果表明:DCGAN可有效生成逼真的螺栓图像,且与常规增广方式相比,DCGAN生成的图像质量更高、性能最优;检测模型受拍摄距离、角度及光照强弱影响且对角度最为敏感,控制拍摄距离在1.6 m内、拍摄角度在20°内、外界光线明亮可保证模型性能较优;与常规增广后训练的模型相比,利用生成图像增广后训练的模型检测性能更优且鲁棒性更好;该智能检测方法可以用于螺栓脱落病害检测,且检测精度明显提高。

基于灰靶决策模型的改性钢渣-沥青混合料路用性能评价

针对钢渣体积稳定性差、钢渣-沥青混合料道路过早开裂的问题,采用二氧化硅胶体溶液对钢渣进行浸泡改性处理,通过力学性能测试、扫描电子显微镜(scanning electron microscope,SEM)检测、路用性能测试等方法研究了改性钢渣的物理力学性能、改性钢渣-沥青混合料的性能和钢渣的改性机理,并引入灰靶理论决策方法,综合改性钢渣-沥青混合料的各项性能指标,确定钢渣的最佳改性方案。结果表明:钢渣改性后,物理力学性能明显提高;钢渣的改性浓度越大,沥青混合料的高温性能越佳;延长钢渣的改性时间,沥青混合料的低温抗裂性能提高;且钢渣改性之后,沥青混合料的水稳定性能显著提高。基于灰靶决策理论,最终确定钢渣的最佳改性方案是在改性浓度(溶液质量分数)为3%的溶液下浸泡24 h。

基于改进YOLOv8n的钢表面缺陷检测模型

针对钢表面缺陷尺寸微小且与背景重合度高而导致的传统缺陷检测算法存在漏检和错检的问题,本文提出了一种基于YOLOv8n改进的钢表面缺陷检测算法。首先,使用全维动态卷积ODConv构建C2f-ODConv模块,在多维度提取钢表面的缺陷特征,增强网络全域特征提取的能力,同时引入基于NWD(Normalized Wasserstein Distance)度量方式的损失函数,提高网络对钢表面缺陷的定位准确度;其次,通过减少网络头部卷积次数,提出新的检测头Light-Detect减少模型的资源占有率,提高模型检测的实时性;最后,在SPPF后加入CBAM注意力机制从而在不同图层上提高特征之间的耦合性。实验结果表明,文中所提算法YOLOv8n-Eff在钢表面缺陷数据集NEU-DET上的mAP(mean Average Precision)值达到78.6%,与YOLOv8n算法相比较,mAP提高了3.2%,计算量减少了2.4 G,并且裂纹目标缺陷的AP值提升了10.8%。结果验证了YOLOv8n-Eff算法可以提升钢表面目标缺陷的平均检测精度,降低漏检率并减少模型计算量,有效满足钢表面缺陷检测的需求。

双碳目标下钢渣固碳的研究进展

在碳达峰、碳中和的背景下,利用钢铁企业废弃钢渣碳酸化固定二氧化碳技术是未来重点研究方向。介绍了中国钢铁行业碳排放现状以及钢渣产生利用的现状,并综述了钢渣直接固碳和间接固碳的研究进展。针对钢渣产生的多元化及大量堆存的问题,提出未来钢渣固碳的研究与发展方向。

“双碳”目标下华菱钢铁财务分析

钢铁行业落实碳达峰及降碳行动是国家完成“双碳”目标的有力支撑,更是实现自身高质量发展的内在需求。财务分析能使企业认识到自身处于何种具体的财务状态,帮助企业提前防范风险,改善经营管理。将从财务指标中的四个能力方面对华菱钢铁的财务报表进行分析,剖析其财务状况,并结合“双碳”目标提出了相关建议。

Calibration of LRFD Format for Steel Jacket OffshorePlatform in China Offshore Area(2): Load, Resistance and Load Combination Factors

Adopting the load and resistance factor design format, the design method for steel jaeket platform structures is developed. Firstly, the limit state equations and design format for steel jacket platform structures are introduced. Then, the ratio of live load effect to dead load effect is estimated. The target reliabilities for design of offshore structures in China offshore area are calibrated by past practice in API RP2A-WSD code. The load and resistance factors are optimized by minimizing the difference within the target reliability and the resulting reliability over the range of load effect ratios. Considering the concurrence of different loads, load combination factors are obtained through an optimization process, and the relation between the load combination factor and load correlation coefficient is established. Finally, the design formulae for steel jacket structures in China offshore area are recommended.

Strain Measurement for Hollow Projectiles During Its Penetration of Concrete Targets

Gives a new technique to measure the dynamic deformation behavior and strain development of a hollow steel projectile during its penetration of concrete targets. Direct strain measurement was performed by applying strain gages attached to the inner walls of the hollow projectile, linked with on-board testing and storage recorder. This on-board test-record system is easy to operate, cost-effective and can provide reasonable, accurate and detailed information. Obverse ballistic experiments were carried out on ogival-nose hollow projectiles normally impacting concrete targets at velocities from 150 m/s to 300 m/s. The deformation process of projectiles was measured, recorded and played back. Profiles of voltage-time relationship were successively obtained and transfered to strain-time relationship with the aid of calibration tables. It was found that projectiles go through a series of compression and tension deformations intermittently. Relationships between strain development and projectile deformation process were discussed.

针对钢材表面小目标缺陷实时检测

钢材表面缺陷检测任务中,YOLO将目标检测转换为对位置信息的回归问题,实现高帧率实时检测,但对小目标缺陷定位精度有所欠缺。针对该问题,以YOLOv5s架构为基础,首先,在模型输入端设定动态尺度训练范式,提高小目标缺陷训练精度;其次,设计STD-CA模块利用图像转换技术,避免下采样过程中分辨率的降低,导致小目标缺陷特征信息的丢失,并引导特征提取能力,降低无关背景特征关注度,进一步提高模型小目标缺陷检测精度。结果表明,在NEU-DET数据集中,改进后模型在保证检测速度保持在54 frame/s的同时,平均精度均值达到86.6%,较YOLOv5s提高17.6%,对小目标缺陷定位更加准确,目前优于其他深度学习钢材实时检测模型。