数字孪生驱动的金属选择性激光熔融成形过程在线监控

由激光熔融过程的极热极冷特性造成的成形零件质量缺陷一直阻碍着选择性激光熔融技术的广泛应用,为制造出高质量的零件,提出一种数字孪生驱动的金属选择性激光熔融成形过程在线监控方法,应用该方法建立智能成形系统并进行案例分析.基于粉末熔融过程的物理和虚拟建模、过程在线监测和大数据分析,构建了包括设备层、数据转换层、网络层、控制层...     

基于深度置信网络的大数据制造过程实时智能监控

针对基于浅层学习模型的过程监控方法难以对大数据制造过程运行状态进行实时智能监控的问题,提出了基于深度置信网络的大数据制造过程实时智能监控方法。利用灰度图建立大数据制造过程质量图谱,以精准表达其过程的运行状态;构建用于识别大数据制造过程质量图谱的深度置信网络;应用离线训练好的深度置信网络模型对当前监控窗口内的过程质量图谱进行识别,实现大数据制造过程实时智能监控。最后,应用该方法对某注塑件大数据制造过程进行实时质量智能监控,结果表明：所提方法的识别性能明显优于基于主成分分析与BP神经网络、支持向量机的识别模型,能有效应用于大数据制造过程实时质量智能监控。     

基于双模式超声换能器的选择性激光熔化316L制件弹性常数无损表征研究

选择性激光熔化技术(SLM)可以快速成型任意结构的金属零件,但其制件的内部组织和材料性能与传统制件有着显著区别。基于双模式超声换能器对SLM 316L不锈钢制件的不同方向弹性常数及其分布进行表征研究。首先通过设计并制备高性能双模式超声换能器,搭建超声弹性常数分布测量系统。对SLM制备的316L不锈钢试样依次进行纵波和两个正交横波声速测量,获得同位置处不同方向弹性模量和泊松比。通过不同成型方向的声速测量,发现SLM制件在堆积z方向熔化层内呈现显著区别,表现出明显的各向异性。而y-z平面内弹性常数分布表征表明,杨氏模量E₁₃大于E₁₂,泊松比σ₁₃小于σ₁₂,且制件各熔化层弹性常数分布规律相似。此外,还讨论了加工过程中扫描速度和扫描间距等工艺参数对弹性参数的影响。选择性激光熔化制件弹性常数的有效无损表征将为增材制件内部质量控制和工艺改善提供技术基础。     

A method for predicting hobbing tool wear based on CNC real-time monitoring data and deep learning

Intelligent monitoring and diagnosis of tool status are of great significance for improving the manufacturing efficiency and accuracy of the workpiece. It is difficult to quickly and accurately predict the wear state of worm gear hob under different working conditions. This paper proposes a novel approach to predict hob wear status based on CNC real-time monitoring data. Based on the open platform communication unified architecture (OPC UA) technology and orthogonal test, the machine data of motor power, current, etc. related to tool wear are collected online in the worm gear machining process. And then, an improved deep belief network (DBN) is used to generate a tool wear model by training data. A growing DBN with transfer learning is introduced to automatically decide its best model structure, which can accelerate its learning process, improve training efficiency and model performance. The experiment results show that the proposed method can effectively predict hob wear status under multi-cutting conditions. To show the advantages of the proposed approach, the performance of the DBN is compared with the traditional back propagation neural network (BP) method in terms of the mean-squared error (MSE). The compared results show that this tool wear prediction method has better prediction accuracy than the traditional BP method during worm gear hobbing.     

基于双隐含层BP算法的激光主动成像识别系统

在传统激光主动成像系统的基础上,结合目标识别技术搭建了一个激光主动成像识别系统实验平台,用于研究激光主动成像后的目标识别。介绍了实验平台的工作原理,基于Hu矩特征的双隐含层BP神经网络算法以及实验处理流程和实验结果。特征量由7个不变Hu矩构成,通过240张原始目标样本库对由136个权值系数构成的双隐含层BP神经网络算法进行了训练。利用训练好的双隐含层BP算法对黑夜条件下远处的运动目标--43式冲锋模具枪进行了实验研究,成功获得了清晰的红外激光主动成像效果。实验显示对450 m处2 740帧和550 m处2 420帧激光主动成像图像的统计识别率达到了68.87%和72.11%,其中旋转变换下的统计识别率可达80.05%和84%,好于仿射变换的识别效果。     

基于SAE-DBN的联合收割机液压系统运行状态监测

液压系统是联合收割机重要的组成系统之一,针对收割机液压系统故障特征提取困难以及多种故障场景下的预警准确率低等问题,提出了一种基于堆叠自编码器和深度信念神经网络融合(SAE-DBN)的联合收割机液压系统运行状态监测方法。在SAE-DBN模型训练过程中,依次训练AE层和RBM层并堆叠,分别得到SAE和DBN,再将SAE和DBN进行连接并微调模型参数。将液压系统中关键部位的参数作为SAE-DBN的输入,进行二次特征提取,然后对液压系统的运行状态进行分类。雷沃GM80型联合收割机的作业运行试验表明:基于SAE-DBN联合收割机液压系统运行状态监测的准确率达到了91.88%,与SAE和BP神经网络等方法相比分别提高了3.82%和8.09%,为液压系统故障诊断提供了一种新的思路。     

激光焊匙孔特征的近红外与X射线传感分析

由于多传感匙孔特征参数可以有效地反映大功率激光焊接质量状态,本文研究了匙孔特征信息的提取方法并建立了焊缝成形预测模型。以大功率盘形激光焊接304不锈钢为试验对象,应用近红外高速摄像机和X射线视觉成像系统同时提取了焊接过程中的熔池动态图像,并分割出匙孔区域。针对近红外图像,应用矩方法导出匙孔的不变矩特征,同时定义并提取匙孔面积和最前端点纵坐标两个特征;针对X射线图像则提取匙孔深度和熵两个特征。在不同激光功率条件下得到匙孔特征并进行特征融合分析,然后建立了3个BP神经网络焊缝成形预测模型。探索了匙孔形态、焊接条件和焊接状态三者之间的联系,实现了对焊接过程的在线监测。试验结果表明,将两个传感器获取的匙孔特征信息融合并进行主成分分析变换后,熔宽和熔深的预测绝对误差平均值分别为0.18mm和0.57mm,比基于单个传感器获取匙孔特征建立的BP神经网络分别减小了0.03mm和0.31mm,显示提出的方法能够有效在线监测大功率盘形激光焊接状态。     

基于神经网络的激光熔覆高度预测

激光成形过程中,对熔覆高度进行实时检测,从而实现熔覆高度闭环控制是成形高质量零件的保证.激光成形过程是一个多参数耦合的非线性过程,大量激光参数对成形熔覆表面质量具有重要影响.在分析激光参数对熔覆高度影响的基础上,建立利用激光工艺参数预测熔覆高度的误差反向传播(Back propagation,BP)神经网络模型,完成了网络算法设计.通过激光成形试验采集样本,利用训练样本对所建立的网络进行训练,完成网络输入输出高度映射关系,并利用测试样本对所训练的网络进行检验.仿真试验表明,神经网络熔覆高度预测模型具有很高的精度,验证了该预测模型在理论和实践上的可行性与有效性.神经网络熔覆高度预测模型为实现激光加工过程熔覆高度实时预测与闭环控制打下基础,对提高成形产品质量具有重要意义.     

基于声发射时频分析与卷积神经网络的液膜密封摩擦状态识别

针对液膜密封状态监测领域无损监测开发不足、信号特征评估困难以及摩擦状态判别智能化特性缺乏的问题,提出一种基于声发射时频分析与卷积神经网络的液膜密封摩擦状态识别方法。该方法将声发射无损监测技术应用于液膜密封的摩擦状态监测,卷积神经网络作为液膜密封摩擦状态自主决策的实现手段,声发射信号的时频信息作为卷积神经网络的特征输入,分析短时傅立叶变换、 S变换以及小波变换3种时频分析方法对卷积神经网络识别性能的影响。结果表明：对于液膜密封的声发射信号,3种时频分析方法与卷积神经网络结合的优选顺序为：短时傅立叶变换、 S变换、小波变换;基于声发射时频分析与卷积神经网络的液膜密封摩擦状态识别方法准确率较高,相比其他识别方法取得了较好的识别效果。     

Application of convolutional neural network to acquisition of clear images for objects with large vertical size in stereo light microscope vision system

For an object with large vertical size that exceeds the certain depth of a stereo light microscope (SLM), its image will be blurred. To obtain clear images, we proposed an image fusion method based on the convolutional neural network (CNN) for the microscopic image sequence. The CNN was designed to discriminate clear and blurred pixels in the source images according to the neighborhood information. To train the CNN, a training set that contained correctly labeled clear and blurred images was created from an open‐access database. The image sequence to be fused was aligned at first. The trained CNN was then used to measure the activity level of each pixel in the aligned source images. The fused image was obtained by taking the pixels with the highest activity levels in the source image sequence. The performance was evaluated using five microscopic image sequences. Compared with other two fusion methods, the proposed method obtained better performance in terms of both visual quality and objective assessment. It is suitable for fusion of the SLM image sequence.     

径向基神经网络在三维激光切割中的应用

分析了RBF神经网络的特点，并研究了它在三维激光切割中的应用，利用该神经网络及其变形可以预测切缝的宽度，也可以根据输入参数对试样的挂渣与否进行判定。比较了三雏激光切割中切割倾角和切割方向对切割质量的影响，从而实现切割质量的预测和优化工艺参数的选取以指导实际的切割过程。     

激光选区熔化单道扫描与搭接数值模拟及试验

激光选区熔化(Selective laser melting, SLM)单道扫描的数值模拟,在数值建模时多采用规则实体模拟铺设的粉末层,通过等效定义材料热物理属性模拟铺设粉末层中粉末与气体共存的情况,难以模拟粉末颗粒的随机性而带来的扫描结果的随机性,且难以分析熔池形貌、内部缺陷的微观衍变过程。针对SLM成形过程中激光功率和单道搭接率对扫描单道和单道搭接质量的影响,以316L不锈钢材料为例,建立SLM首层单道扫描与单道搭接数值模拟模型。SLM的铺粉过程在基于离散单元法的EDEM中建立,并得到数值化的粉床几何模型。SLM的单道扫描与搭接的模拟基于有限体积法,在FLUENT中实现,采用两相流模型与熔化/凝固模型捕捉熔池形貌变化过程,得到不同激光功率和扫描搭接率下成形单道与单道搭接的数值模型。最后结合试验表明了在100～300 W激光功率下成形单道表面形貌与缺陷的形成,当激光功率为100～150 W时,单道形貌不规则,且容易形成局部缺陷;在200～300W功率下,激光功率越大,保证搭接质量的最低搭接率越小,当激光功率为250 W时,应保证单道填充间距不大于0.1 mm。研究成果对SLM工艺参数...     

基于小波神经网络和证据理论的零件图像识别

提出基本概率分配构造和多源零件图像特征识别的方法。首先,获取零件图像的小波分解系数和零件图像的相对边缘像素系数。然后,以零件图像的小波分解系数和零件图像的相对边缘像素系数作为零件图像的特征,并作为神经网络的输入,得到多源零件图像识别的基本概率分配。最后,依据证据理论的合成规则得到零件的识别结果。实验结果表明,文中提出的方法是有效的。     

Data mining and statistical inference in selective laser melting

Selective laser melting (SLM) is an additive manufacturing process that builds a complex three-dimensional part, layer-by-layer, using a laser beam to fuse fine metal powder together. The design freedom afforded by SLM comes associated with complexity. As the physical phenomena occur over a broad range of length and time scales, the computational cost of modeling the process is high. At the same time, the large number of parameters that control the quality of a part make experiments expensive. In this paper, we describe ways in which we can use data mining and statistical inference techniques to intelligently combine simulations and experiments to build parts with desired properties. We start with a brief summary of prior work in finding process parameters for high-density parts. We then expand on this work to show how we can improve the approach by using feature selection techniques to identify important variables, data-driven surrogate models to reduce computational costs, improved sampling techniques to cover the design space adequately, and uncertainty analysis for statistical inference. Our results indicate that techniques from data mining and statistics can complement those from physical modeling to provide greater insight into complex processes such as selective laser melting.     

基于SOM神经网络的摩擦状态识别研究

为提取摩擦振动的特征和实现摩擦副摩擦状态的识别,在往复摩擦磨损试验机进行摩擦副混合摩擦和干摩擦状态的摩擦磨损试验。应用谱减法对试验采集的摩擦振动信号进行降噪,计算降噪后的摩擦振动15个特征参数。应用自组织映射(Self-organizing map, SOM)神经网络对摩擦副不同摩擦状态的摩擦振动特征参数进行分析,得到摩擦振动的SOM神经网络神经元分类。研究结果表明,谱减法能消除摩擦磨损试验机的背景噪声,SOM神经网络算法能够有效分析摩擦振动信号的特征,实现摩擦副摩擦状态的识别。     

基于神经网络的污水处理设备监测CPS设计

现阶段大部分污水处理设备使用单一传感器监测设备,故障识别率低。该文研发了一套新型污水处理设备监测系统,根据信息物理融合系统(CPS)的要求全面统筹,采用BP神经网络算法处理由传感器得到的数据,监测设备的不同状态。硬件芯片采用STM32,采用工业以太网和串口通信,采用C#语言开发上位机应用软件,采用HTML和JavaScript以及C#语言开发web远程监测系统。经测试,系统运行良好,故障识别率高。     

基于神经网络算法的故障检测技术

针对复杂的机电产品内部构件状态检测这一工程难题,本文介绍了一种自动在线检测系统.该系统采用X射线对产品成像,运用数字图像处理技术对射线图像进行预处理,由神经网络算法进行故障诊断.故障识别模型采用了改进的BP神经网络算法,以正常装配状态时的多幅图像经预处理后作为学习样本训练BP神经网络.检测时一般只需拍摄两幅不同方位的图像,经预处理后输入神经网络与样本图像进行比较判断,即可识别出关键元器件的状态.该系统将数字射线成像技术和图像处理技术相结合,并在故障识别算法中采用了神经网络算法,提高了产品故障的检测速度和可靠性,在工业无损检测领域具有一定的实用性.     

Formulation of bead width model of an SLM prototype using modified multi-gene genetic programming approach

The rapid prototyping (RP) processes, specifically selective laser melting (SLM), are popular for building complex 3D parts directly from the metal powder. The literature reveals that the properties such as surface roughness, waviness, bead width, compressive strength, tensile strength, wear, and dimensional accuracy of an SLM-fabricated prototype depend on the parameter settings of the SLM setup and can be improved by appropriate adjustment. For the selection of an optimal parameter setting, multi-gene genetic programming (MGGP), which develops the model structure and its coefficients automatically, can be applied. The model participating in the evolutionary stage of the MGGP method is a linear weighted sum of several genes (model trees) regressed using the least squares method. In this combination mechanism, the occurrence of gene of lower performance in the MGGP model can degrade its performance. This paper proposes a modified MGGP (M-MGGP) method using a stepwise regression approach such that the genes of lower performance are eliminated, and only the high performing genes are combined. The M-MGGP approach is applied on the bead width data obtained from the experiments conducted on the SLM machine, and its performance is found to be better than that of the standardized MGGP and artificial neural network (ANN) models. Between MGGP and ANN, ANN has shown better performance. Further, the parametric and sensitivity analysis conducted validates the robustness of our proposed model and is proved to capture the dynamics of the SLM process by unveiling important process parameters and the hidden non-linear relationships.     

磨削加工参数智能化在线调整方法研究

提出运用神经网络的非线性建模原理和自学习能力，建立影响磨削参数选择的各种因素与磨削参数之间的函数映射关系。据此可以实现初始磨削用量智能化选择目的。将磨削过程智能监测模块检测到的砂轮状态特征信号反馈到神经网络的输入端，根据砂轮钝化程度实现磨削参数的在线智能调整，以减少砂轮钝化对磨削质量的影响，确保加工质量的稳定性。     

蜗轮减速器振动信号特征提取与状态识别

为实现蜗轮减速器运行状态识别,首先结合小波包分解和矩阵理论的特点,提出基于参考信号的小波包能量矩阵构造方法,分析了矩阵的最大奇异值（特征值）与运行状态的物理联系,并验证了所提方法比以往方法提取出的特征参数敏感度更高;然后改进思维进化算法（MEA）用于优化BP神经网络,实现对运行状态的智能识别,将提取的特征参数构成神经网络的输入向量,结果表明识别正确率提高了17.93%,从而验证了改进算法的优越性;最后提出了一种快速分类方法,该方法可以较好地区分故障与正常状态,解决了对实时性要求较高的在线诊断问题。