基于回归分析理论的点焊质量监测模型

以动态电阻与接头质量参数之间关系为例,用线性回归分析技术,系统地研究了点焊加热过程中的监测信息与点焊接头质量参数之间的一元和多元回归模型关系。结果表明：（1）各种回归模型总体线性关系经统计检验均为高显著,可以用于点焊接头质量监测;（2）一元回归模型的决定系数仅为0．15－0．33,多元回归模型的决定系数则可以提高到0．59－0．79,说明发展多参量综合监测技术是提高点焊质量监控精度的有效途径;（3）点焊过程监测信息与质量参数之间关系也含有较大的非线性,用线性模型去描述这样的关系导致模型误差的增加。     

Effects of modeling means on properties of monitoring models of spot welding quality

Analyzing and modeling the relation between monitoring information during welding and quality information of the joints is the foundation of monitoring resistance spot welding quality. According to the means of modeling, the known models can be divided into three large categories: single linear regression models, multiple linear regression models and multiple non-linear models. By modeling the relations between dynamic resistance information and welding quality parameters with different means, this paper analyzes effects of modeling means on performances of monitoring models of resistance spot welding quality. From the test results, the following conclusions can be drawn: By comparison with two other kinds of models, artificial neural network (ANN) model can describe non-linear and high coupling relationship between monitoring information and quality information more reasonably, improve performance of monitoring model remarkably, and make the estimated values of welding quality parameters more accurate and reliable.     

运动学误差模型变体下的机械臂标定精度研究

为进一步提高机械臂运动学标定精度,研究不同D-H建模方式下运动学误差辨识精度的差异,为运动学误差准确建模提供理论依据。围绕SD-H和MD-H运动学误差模型的各种变体形式,量化对比各模型变体的性态水平和线性化误差,并分析二者对运动学标定精度的综合影响。研究结果表明：x-z-y欧拉变换会破坏MD-H运动学误差模型最小性,不适用于运动学标定;其余满秩MD-H模型性态水平优于SD-H模型两个数量级;MD-H建模时引入的线性化误差与SD-H数量级等同;在模型性态水平和线性化误差综合作用下,较强测量噪声环境中的SD-H模型运动学标定精度(0.1%~39.3%)优于MD-H模型(0.1%~71.2%);角度类与位置类的运动学参数误差的辨识精度差异显著。研究成果为机械臂标定的精确建模以及误差类型与辨识方法的合理匹配提供了理论依据。     

减小点焊质量神经网络监测模型误差的措施

点焊过程监测信息与质量参数之间关系也含有较大的非线性,用线性模型去描述这样的关系将导致模型误差的增加。为了更好地描述点焊过程监测信息与质量参数之间的复杂关系,文中将神经网络理论用于点焊过程模型化。在建立点焊质量神经网络监测模型的过程中,发现,训练过程中的“假饱和”现象是减小网络模型误差的主要障碍。为此,分析了各种减小网络模型误差的可能途径,提出了相应的改善措施,并通过试验证明,所提出的观点是正确的、措施是行之有效的。     

Comparison of two analytical models of blanking and proposal of a new model

With the development of accurate Finite Element models of the punching/blanking process by the present authors and other researchers, important possibilities emerged to model and analyse the process in a research environment. However, Finite Element models are less well suited for use in a production environment, to monitor the process in real time. For real-time monitoring applications, an analytical model could potentially be used to extract relevant information from measurements of force and displacement.In this article, first, two existing analytical models of blanking are compared. One model is based on the assumption of pure shear and the other is a tension model. Secondly, an addition to the shear model is proposed, resulting in a new analytical model, which allows for the modelling of plastic bending during the process. Results of the new model are compared against experimental findings, and show good similarity for the materials tested. Particular attention is paid to the assumptions and the stress–strain systems inherent in the models.     

工件同轴度误差的视觉检测系统

建立了一种基于虚拟仪器技术的工件同轴度误差检测系统.该系统选用蓝色平行光作为背光源,经过中值滤波、图像二值化、模板匹配、边缘提取等图像处理技术,采用最小二乘法拟合轴心线来计算工件的同轴度误差,同时提出了自动迭代最优化搜索阈值法和环境变化实时监控方法.实验结果表明,此系统检测结果准确可靠,具有良好的实用价值.     

小功率激光模具自动修复CCD标定技术

针对模具修复过程中的质量控制问题引入视觉传感技术. 为得到准确的模具修复中的视觉特征信息,对模具修复试验装置的成像系统进行了分析,建立成像模型,分析亚像素角点提取算法,制作符合试验装置要求的标定板,对利用改进的张正友法对试验装置进行了标定,求得成像系统的内参数和外参数. 结果表明,成像系统标定误差为0.712 305像素,即标定误差为0.012 4 mm,利用标定后的试验装置修复模具;修复结果表明标定方法实用可靠,满足小功率激光模具自动修复的精度要求.     

Development of a tool wear-monitoring system for hard turning

This paper describes an in-depth study on the development of a system for monitoring tool wear in hard turning. Hard turning is used in the manufacturing industry as an economic alternative to grinding, but the reliability of hard turning processes is often unpredictable. One of the main factors affecting the reliability of hard turning is tool wear. Conventional wear-monitoring systems for turning operations cannot be used for monitoring tools used in hard turning because a conglomeration of phenomena, such as chip formation, tool wear and surface finish during hard turning, exhibits unique behavior not found in regular turning operations. In this study, various aspects associated with hard turning were investigated with the aim of designing an accurate tool wear-monitoring system for hard turning. The findings of the investigation showed that the best method to monitor tool wear during hard turning would be by means of force-based monitoring with an Artificial Intelligence (AI) model. The novel formulation of the proposed AI model enables it to provide an accurate solution for monitoring crater and flank wear during hard turning. The suggested wear-monitoring system is simple and flexible enough for online implementation, which will allow more reliable hard turning in industry.     

麻醉相关神经网络在麻醉深度监测中的研究进展

麻醉深度把握不当不仅不利于麻醉的快速平稳复苏,也可影响患者术后转归。因此,精准把控麻醉深度是麻醉学领域亟待解决的临床和科学问题。目前临床主要基于脑电图（EEG）信号监测衍生出的不同算法模型进行麻醉深度的监测,然而并不能满足麻醉医生对于麻醉深度精准把握的要求。近年来,有关麻醉相关神经网络机制及其调制的研究提示,其作为麻醉深度监测的方法具有潜在价值。麻醉相关神经网络主要包括睡眠觉醒环路、丘脑-皮质环路及皮质-皮质网络。深入理解参与麻醉引起意识消失的神经网络,将更为精准地指导麻醉深度监测,为提高临床麻醉复苏质量提供可能。     

基于Faster R-CNN神经网络的钢卷端部缺陷检测

钢卷的端部质量不但影响后续生产的效率和稳定性,还影响产品的整体产品质量.然而目前国内外大部分钢铁企业仍是靠人工进行检测,尚未有高效、准确的设备对端部质量进行监控.为此,研究了通过机器视觉和Faster R-CNN神经网络对钢卷端部缺陷进行定位和检测的方法.研究发现,Faster R-CNN神经网络在经过足量的有效数据训...     

Admissibility of Bayes estimate with inaccurate prior in surveying adjustment

1　ＩＮＴＲＯＤＵＣＴＩＯＮＢａｙｅｓｅｓｔｉｍａｔｅｈａｓｂｅｅｎｗｉｄｅｌｙｕｓｅｄｉｎｓｕｒｖｅｙｉｎｇｄａｔａｐｒｏｃｅｓｓｉｎｇ[1～ 3].ＴｈｅａｄｖａｎｔａｇｅｏｆＢａｙｅｓｅｓｔｉ ｍａｔｅｉｓｔｈａｔｉｔｃａｎｍａｋｅｕｓｅｏｆｎｏｔｏｎｌｙｔｈｅｉｎｆｏｒｍａ ｔｉｏｎｃｏｎｔａｉｎｅｄｉｎｔｈｅｏｂｓｅｒｖａｔｉｏｎｓ,ｂｕｔａｌｓｏｔｈｅｉｎｆｏｒ ｍａｔｉｏｎｆｒｏｍｔｈｅｈｉｓｔｏｒｉｃａｌｄａｔａｏｒｆｒｏｍｏｔｈｅｒａｒｅａｓ.ＴｈｉｓｍｅａｎｓｔｈａｔｔｈｅＢａｙｅｓｅｓ…     

基于光栅投影的焊后焊缝表面三维测量

焊缝表面三维轮廓是评价焊后焊缝质量的重要指标,针对常用线结构光扫描进行焊缝测量时无法兼顾测量速度与精度等问题,设计并搭建了一种基于面结构光光栅投影的焊缝三维轮廓测量系统. 首先,通过数字光处理(digital light processing,DLP)投影仪向焊缝表面投射面结构光光栅条纹图像,摄像机获取变形调制条纹,利用四步相移法结合多频外差时域解相算法准确地实现了变形光栅条纹相位主值的解算和相位的展开.然后,利用平面靶标结合精密平移台获取空间点阵列的方法实现了对模型参数的标定,该标定方法精度高、可操作性强、结构稳定.最后,采用空间相位映射模型实现焊缝相位信息到高度信息的转换,实现焊缝三维轮廓的测量.结果表明,该测量方法能很好的表现焊缝细节信息,测量结果准确,测量精度能达到0.0968 mm,可以为焊后焊缝外观检测与评价提供可靠数据.     

飞机载荷机构自动测试系统设计

对飞机载荷机构性能及测试工艺进行分析,运用现代检测理论与技术,设计了飞机载荷机构自动测试系统的硬件结构和软件结构,实现了对载荷机构性能的自动检测与判断,解决了机务部队无法对其性能进行检测的难题。经使用表明:该系统稳定可靠、检测精度高、使用维护方便、拓展性好,满足了机务部队对各型载荷机构性能检测的要求。     

数控机床的热误差建模与补偿研究

数控机床的热变形误差是影响其加工精度的主要因素。针对当前机床热误差难以解决的问题,提出一种融合模糊聚类理论、灰色关联理论和多元线性回归理论的热误差建模及补偿原理,通过应用于实验室自主研制的AGPM,经机床温度场的测点优化分析、多元线性回归求解,建立了精确的热误差补偿模型。经补偿验证,该原理理论正确、简单易行、稳定可靠,可以大幅减小机床的热变形误差。     

点焊质量监控技术的发展与现状

根据控制目标的不同，将已有的点焊质量监控方法分为四类。通过对各类方法进行详细分析，认为发展多参量综合监测技术是提高点焊质量监督控精度的有效途径。其中，如何充分利用监测信息，采用合理的建模手段，建立合理的多元非线性监测模型，并使监测模型能够在较宽条件内提供准确、可靠的点焊质量信息，则是发展多参量综合监控技术的关键。利用神经元网络理论和模糊逻辑理论，可望解决点焊质量控制领域中许多长期以来难以解决的问题，有利于实现真正的点焊质量直接控制。     

基于贝叶斯图像模式识别技术的点焊质量评估

提出一种将点焊过程动态电极位移信号转化为二值图像的方法.基于图像特征分析,从焊点样本电极位移二值图像中提取出15个隐含特征.针对一系列对应不同焊接质量焊点样本电极位移二值图像特征,利用主成分分析消除图像特征间的互相关性,建立了基于最小风险贝叶斯图像识别技术的焊点质量分类器.分类器有效性测试结果表明,电极位移信号二值图像尽可能多的保留了焊点质量信息,特征提取算法简单、高效、易于实现;同时在小数据样本情况下,贝叶斯图像识别技术能够快速、准确地评判焊点质量,有较好的应用前景.     

基于信度评价的焊缝结构光自适应多次曝光法

利用线结构光主动视觉法获取焊缝尺寸时,被测工件表面的复杂特性会造成局部图像质量下降,严重干扰焊缝尺寸计算.针对此问题提出了一种基于信度评价的自适应多次曝光法.将光条的信度评价作为结构光图像的质量评价参数,根据光条的信度分布自动获得下一次拍摄的曝光参数并判断拍摄结束节点.结果表明,基于信度评价的焊缝结构光自适应多次曝光法应用于角焊缝尺寸检测,具有良好的测量稳定性与测量精确度,并且可以有效避免光条图像质量退化造成的测量误差和测量数据缺失.     

Straightness error compensation for large CNC gantry type milling centers based on B-spline curves modeling

Fast and accurate modeling for the straightness errors of machine tools is significant important for the implementation of error compensation. To decrease the straightness errors which generally exist in gantry type milling centers, a novel approach for real-time compensation of straightness errors is presented, including an accurate spatial straightness error model which based on B-Spline curves method and a real-time errors compensation system which can compensate multiple errors at the same time in real-time. An experiment is carried out to measure the straightness errors of a gantry type milling center by using a laser interferometer. There are two straightness errors for one direction movement of an axis. As a result, there are six straightness errors for a three-axis milling center. The straightness error model is established by the B-Spline curves method. According to these six straightness error models, a spatial straightness error model is obtained by superposition of these straightness error models. In addition, an external real-time errors compensation system is developed based on the function of external mechanical origin offset in Fanuc CNC systems. The compensation experiments are conducted to verify the accuracy of the spatial straightness error model and the effectiveness of the error compensation system. The experimental results show that the straightness errors of the machined work-pieces are proved to have been decreased over 90% compared with that of non-compensations. The B-Spline curves modeling combining the errors compensation system can be utilized as an effective approach to improve the accuracy of the CNC gantry type milling centers.     

A systematic approach to enhance machine tool accuracy for precision manufacturing

As competitive pressures demand higher levels of productivity and automation, accuracy and reliability of products become increasingly important. Innovations in manufacturing technology also are driven by demands to shorten production cycle time and to maintain a consistently high level of product quality in an advanced manufacturing environment. Innovative and cost-effective methods must be developed to advance precision manufacturing technologies to compete in the global economy. In this paper, a systematic approach is presented to advance the accuracy and functionality of multi-axis machines for precision manufacturing. Four major steps are introduced. First, a method is presented for rapid characterizing of the accuracy of multi-axis machines. Detailed derivation of a generalized error model is given. Second, a method is discussed on the feasible applications of on-machine sensing and analysis. The information obtained from on-machine sensing process is used with the information collected from post-process inspection to improve the machine performance. Third, a method is presented for utilizing post-process inspection and innovative data analysis techniques to decouple the effects of process- and machine-related errors. Finally, an adaptive error correction system is outlined for iterative identification of the residual error and the fine-tuning of the error model for more accurate error modeling and compensation.     

Reliable multi-objective optimization of high-speed WEDM process based on Gaussian process regression

The paper discusses the development of reliable multi-objective optimization based on Gaussian process regression (GPR) to optimize the high-speed wire-cut electrical discharge machining (WEDM-HS) process, considering mean current, on-time and off-time as input features and material remove rate (MRR) and Surface Roughness (SR) as output responses. In order to achieve an accurate estimation for the nonlinear electrical discharging and thermal erosion process, the multiple GPR models due to its simplicity and flexibility identify WEDM-HS process with measurement noise. Objective functions of predictive reliability multi-objectives optimization are built by probabilistic variance of predictive response used as empirical reliability measurement and responses of GPR models. Finally, the cluster class centers of Pareto front are the optional solutions to be chosen. Experiments on WEDM-HS (DK7732C2) are conducted to evaluate the proposed intelligent approach in terms of optimization process accuracy and reliability. The experimental result shows that GPR models have the advantage over other regressive models in terms of model accuracy and feature scaling and probabilistic variance. Given the regulable coefficient parameters, the experimental optimization and optional solutions show the effectiveness of controlling optimization process to acquire more reliable optimum predictive solutions.