工序能力指数计算中标准差估计的讨论

计算工序能力指数时,通常采用样本级差或样本总体标准差估计总体标准差.在实际机械加工过程中由于刀具磨损等因素导致加工的工件尺寸发生漂移,需要进行定期的调整,根据实际情况比较了不同标准差估计方法所计算的工序能力指数,分析了原因,并通过仿真进行了研究.     

自相关过程能力分析

在过程能力分析中,过程能力指数计算通常采用样本级差或者样本标准差来估计过程输出的标准差.实际制造过程中,一些过程输出存在自相关特性.在时间序列理论中的稳定随机过程模型的基础上,通过仿真的方法研究了过程输出存在自相关特性时,传统统计过程控制抽样和标准差估计方法以及样本含量对过程能力指数计算的影响,指出必须考虑过程输出的相关性,并通过实例进行了说明.     

顺序迭代开发过程建模与仿真

为有效地管理和改进产品开发过程,提出了一种基于奖赏马尔科夫链的顺序迭代过程定量模型.该模型与已有的顺序迭代过程模型不同的是,模型中的返工影响因子不仅作用于直接返工任务的持续时间,而且作用于该返工任务的后续任务,即作用于返工任务的整个剩余时间.一个任务在每个阶段的返工量随着迭代次数的增加而逐渐减少.考虑了不同任务引起的同一任务返工量的不同.为估计开发过程的期望时间,建立了过程时间估计的分析模型和仿真模型,给出了仿真计算算法.以软件测试过程为例,给出了过程时间估计的分析和仿真实验结果,并与其他已有模型进行了对比分析.最后,讨论了模型估计结果产生偏差的可能原因.     

卡尔曼滤波器在过程估计中的应用

提出了一种卡尔曼滤波算法,建立离散状态模型,运用MATLAB对直流电机进行分析,对过程进行估计.以一个平方信号作为过程输入,分别计算了无噪声过程响应和有噪声过程响应来加以比较.根据控制信号和测量输出估计过程输出和状态,卡尔曼滤波器的估计输出接近于无噪声过程的输出,状态估计也令人满意;估计输出信号的误差方差较小,效果良好.     

相关系数平稳过程方法

工程实际中遇到的随机过程,其均值和方差多数都随时间而变化,因此,无法用传统的平稳随机过程理论描述。文中通过对大量随机过程实测数据的分析研究,发现它们的相关系数函数基本上不随时间平移而改变,即相关系数函数是平稳的。在此基础上建立一种新的相关系数平稳随机过程理论和方法,并给出相关系数平稳正态过程的极大似然估计。从而在性能测试中,采用本文方法只要测定一个样本函数就能够对产品性能的均值和方差进行估计,与传统的需要测定一组样本函数的成组试验法相比,可以节省大量试验。     

产品生产过程能力综合评价模型研究

过程能力指数是一种广泛应用的产品质量评价工具,现有的过程能力评价方法只针对加工过程,对于生产全过程的能力综合评价缺乏应有的方法.本文提出了一种充分利用生产各阶段检验数据的产品生产过程能力综合评价模型.该模型将加工过程的能力评价方法扩展到人厂检验、装配过程和出厂检验,提出了生产过程能力评价指标体系,并采用层次分析法确定各自权重,通过加权求和得出生产过程能力总指数.最后通过一个实例验证了该评价模型可以客观量化生产过程的总体能力,该模型对于监控和改进生产过程能力有重要意义.     

正态疲劳母体样本标准差的Weibull分布特征及应用

采用数值方法,研究了服从正态分布的总体在不同样本容量时其样本标准差的分布特征及变化规律,结果表明小子样条件下样本标准差呈偏态分布,具有Weibull分布的特征。利用三参数Weibull分布归纳出分布参数的数学模型,进而提出了一种具有更高精度的小子样总体标准差区间估计方法,并得到试验数据的验证。     

基于过程信息统计分析的微点焊过程监控

为加强对漆包线-铜箔微点焊过程的监测,便于管控实际焊接生产质量,构建了微点焊过程监控平台。基于过程电信息数据,采用最小二乘法对微点焊电极全生命周期动态电阻变化规律进行建模,确定电极烧损状态变化临界点,实现电极烧损状态监测;利用统计过程控制技术,通过生产过程电信号均值-标准差图,实现微点焊过程异常信息监测与质量管控。     

线性过程回归分析

提出线性过程的定义及其回归分析方法,建立线性过程回归模型、回归方程和高置信水平、高可靠度的置信上、下限曲线。线性过程是具有线性函数均值和二次函数方差的独立增量随机过程,工程中的蠕变、松弛、刚度降、裂纹扩展、性能退化等许多随机过程均可变换为线性过程。传统回归分析只能对数据点进行处理,文中方法则能对过程进行回归分析,可以将线性过程作为一个整体进行统计推断,充分开发利用不同时刻试验数据之间的纵向信息,增大了信息量。从而在精度相同的情况下可节省大量试样,在试验量相同的情况下则能显著提高估计精度。     

自适应制造过程的神经网络控制

分析神经网络应用于自适应控制的有效途径。提出一种基于过程增益估计的神经网络特殊学习自适应控制方案,并实现对时变加工过程的有效控制。     

A transformation technique to estimate the process capability index for non-normal processes

Estimating the process capability index (PCI) for non-normal processes has been discussed by many researches. There are two basic approaches to estimating the PCI for non-normal processes. The first commonly used approach is to transform the non-normal data into normal data using transformation techniques and then use a conventional normal method to estimate the PCI for transformed data. This is a straightforward approach and is easy to deploy. The alternate approach is to use non-normal percentiles to calculate the PCI. The latter approach is not easy to implement and a deviation in estimating the distribution of the process may affect the efficacy of the estimated PCI. The aim of this paper is to estimate the PCI for non-normal processes using a transformation technique called root transformation. The efficacy of the proposed technique is assessed by conducting a simulation study using gamma, Weibull, and beta distributions. The root transformation technique is used to estimate the PCI for each set of simulated data. These results are then compared with the PCI obtained using exact percentiles and the Box-Cox method. Finally, a case study based on real-world data is presented.     

Value of information-based experimental design: Application to process damping in milling

This paper describes a value of information-based experimental design method that uses Bayesian inference for belief updating. The application is process damping coefficient identification in milling. An analytical process damping algorithm is used to model the prior distribution of the stability boundary (between stable and unstable cutting conditions). The prior distribution is updated using experimental results via Bayesian inference. The updated distribution of the stability boundary is used to determine the posterior process damping coefficient value. A value of information approach for experimental test point selection is then demonstrated which minimizes the number of experiments required to determine the process damping coefficient. Subsequent experimental parameters are selected such that the percent reduction in the standard deviation of the process damping coefficient is maximized. The method is validated by comparing the process damping posterior values to residual sum of squares results using a grid-based experimental design approach. Results show a significant reduction in the number of experiments required for process damping coefficient parameter determination. The advantages of using the value of information approach over the traditional design of experimental methods are discussed.     

The measurement of a process capability for folded normal process data

There are two approaches to estimate the nonconforming percentage and process capability indices which are commonly used to measure the ability of a process to manufacture products meeting a required specification. The most commonly used approaches are based on the assumption that the process is normally distributed. If the process is not normal with a skewed distribution, but normal-based techniques are used to estimate process capability indices or fraction nonconforming, serious errors can result. Basically, non-normal process data, such as folded normal process data, is common in mechanical industries. In this paper, we propose one procedure to evaluate the process capability from folded normal data, by solving the true process mean and process standard deviation of the underlying normal distribution. A numerical result based on the proposed approach is illustrated.     

Calculating process control parameters from steady state operating data

The tutorial paper will show that chemical process control parameters can be obtained from steady state operating data in combination with simple algebraic equations. The parameters include steady state process gain, feedforward control factor and approximate process time constant. Here, we treat ‘self-regulating’ processes first. Integrating processes and unstable processes are treated separately, unlike the standard method of setting up linear differential equations followed by Laplace transformations. With the standard method, a steady state portion of dynamic response, if there is any, appears as a part of equations of a complex variable. Many control engineers in the chemical industries who might have studied the method, do not use it because of mathematical difficulties. Instead, they may resort to plant tests that are often time consuming and costly. The parameters obtained by this paper's method will aid improvement of control systems. We will address the relation of this method with that of commercial software, which implement advanced process controls based on plant tests.     

Comparing four lower confidence limits for process yield index Spk

Process yield has long been a standard criterion used in the manufacturing industry as a common measure on process performance. Boyles considered a yield index S_pk for normal processes. The measurement S_pk establishes the relationship between the manufacturing specifications and the actual process performance, which provides an exact measure on the process yield. In this paper, the natural estimator of S_pk is considered using the bootstrap simulation technique to find four approximate lower confidence limits. The four bootstrap methods including the standard bootstrap (SB), the percentile bootstrap (PB), the biased corrected percentile bootstrap (BCPB), and the bias-corrected and accelerated (BCa) bootstrap methods are compared based on the coverage fraction. The simulation results show that the SB method significantly outperforms PB, BCPB and BCa, and therefore is recommended for use in assessing process performance S_pk based on the yield.     

Estimating process capability indices of multivariate nonnormal processes

The capability analysis of production processes where there are more than one correlated quality variables is a complicated task. The problem becomes even more difficult when these variables exhibit nonnormal characteristics. In this paper, a new methodology is proposed to estimate process capability indices (PCIs) of multivariate nonnormal processes. In the proposed methodology, the skewness of the marginal probability distributions of the variables is first diminished by a root transformation technique. Then, a Monte Carlo simulation method is employed to estimate the process proportion of nonconformities (PNC). Next, the relationship between PNC and PCI is found, and finally, PCI is estimated using PNC. Several multivariate nonnormal distributions such as Beta, Weibull, and Gamma are taken into account in simulation experiments. A real-world problem is also given to demonstrate the application of the proposed procedure. The results obtained from both the simulation studies and the real-world problem show that the proposed method performs well and is able to estimate PCI properly.     

Determining the optimum process mean for a mixed quality loss function

In 1999, Wen and Mergen adopted the step loss function of a product to balance the costs of out-of-specification occurrences and to determine the optimum process mean. They considered a normal quality characteristic and assumed a known standard deviation for the process in their model. This paper proposes a modified Wen and Mergen [1] model with a mixed quality loss function to determine the optimum process mean. The quality loss of the product per item includes a quadratic loss within specifications and an out-of-specification piecewise linear loss.     

基于12维统计矢量的GMAW焊接过程监测模糊神经网络系统

将熔化极气体保护焊(GMAW)焊接电参数概率密度分布(PDD)和时间频数分布(CFD)数值信息进行进一步的处理,用其平均值、方差和标准方差等统计参数,构成12维矢量S12,描述不同工艺条件下的GMAW焊接过程。综合神经网络和模糊技术的优点,建立了模糊神经网络系统FKCN,对8种工艺条件下24个GMAW焊接试验的识别成功率达到了100％。     

基于数理统计的零件热处理工序能力分析

为了分析小销轴失效的原因,对小销轴的硬度抽样检测数据进行了数理统计分析,计算了零件淬火热处理工序的工艺能力系数。分析结果表明,虽然抽样检测数据都在公差范围之内,但该分布的标准差偏大,工序能力系数小于1,说明该热处理工序的工艺能力不足,必然会导致不合格品的出现。对不合格率进行了估算。结果表明,该工序出现不合格品的概率为1．17％。工艺能力不足是导致部分小销轴失效的根本原因。     

A synthetic control chart for monitoring process dispersion with sample range

A synthetic control chart for monitoring the changes in the standard deviation of a normally distributed process is proposed in this paper. The synthetic chart consists of the sample range (R) chart and the conforming run-length (CRL) chart. The R chart can be viewed as a special case of the synthetic chart. The operation, design and performance of this chart are described. Average run- length comparisons between other procedures and the synthetic chart are presented. It indicates that the synthetic chart is a good alternative for monitoring process dispersion. The variable sampling interval (VSI) schemes, as an enhancement to the synthetic chart, are discussed to further improve the chart performance. An example is presented to illustrate the application of synthetic chart and its VSI scheme.