中文式质量管理论文.docx

1、中文式质量管理论文通过设计和实验确定最佳质量控制系统摘要不是所有的方法都是同样有效的质量控制。在实验室测试不合格中，基本上是由过度的工艺变化和失误导致的。统计质量控制可以有效地控制过程的变化，但它却无法检测或防止大多数的错误。由于错误或失误是经常导致不合格的主要来源，我们得出统计质量控制本身是不是有效的。我探索100%检验方法控制错误。不同的检测技术，戴明描述为无效的，新的“源”的检测方法可以检测错误，使在不合格产生前修正，在传统方法的成本中一小部分达到质量最高程度。描述了任务的复杂性与不合格率之间的重要关系，随着文化的变化，实现最佳的质量控制的做法是必不可少的。关键词：防错 统计质量控制 任

2、务复杂度 缺陷 不合格最近，通过组织活检我的一个朋友被确定她的乳腺癌已经扩散到了淋巴结。淋巴结手术室和实验室之间没有淋巴结，她现在忍受痛苦的化疗即使可能没有乳腺癌转移。这样的错误是在几乎每一个行业存在。这个案例说明，错误往往导致生产力的损失和浪费，那些我们努力去改正。错误无法用传统的方法如统计质量控制（QC）控制，一个对质量控制更严格的方法被需要。To understand the best methods for eliminating nonconformities in any process, one must first understand the sources of nonco

3、nformities and the most effective ways to eliminate or control each type. Toward this goal I review the evolution of quality control and modern quality-control methods that are achieving revolutionary reductions in defect rates.了解消除任何过程不合格最好的方法，首先必须了解不合格的来源和最有效的方法来消除或控制的每个类型。朝着这个目标我审查减少缺陷率的质量控制和质量控制

4、方法。传统的质量控制一个质量体系每一个操作或生产系统目标是生产一个有用的产品。产品可能是一种服务，信息，或物理对象。每个生产周期的开始输入，是由一个过程转变成一个更理想的状态或产品。（1）传统生产的投入到每一个过程：人（执行和控制过程的人）；机（用于执行处理的设备或机械）；材料（工艺要求原材料或零部件）方法（用于执行程序和序列）信息（工作指令，数据，和传感器的读数，引导过程执行）。In each process, excessive variations and errors can cause nonconformities, with three undesirable consequen

5、ces: (a) scrapped or wasted resources; (b) degraded process throughput; (c) “contamination” from undetected nonconformities, reducing the value of the product to the customer. 在每个过程中，过多的变化和错误可能会导致不合格，有三的不良后果：（a）报废或资源的浪费；（b）降解过程吞吐量；（c）发现的不符合项“污染”，降低产品的价值给客户。The goal of quality control in every produc

6、tion system is to (a) eliminate nonconformities and their consequences, (b) eliminate rework and wasted resources, and (c) achieve these goals at the lowest possible cost.每一个生产系统质量控制的目标是（）消除不合格及其后果，（b）消除返工和浪费资源，（c）尽可能低的成本实现这些目标质量控制的发展大多数的质量控制方法，最初开发为了辅助制造。这并不奇怪，因为高量的生产通常需要多次重复涉及控制操作顺序。操作经常重复，更容易识别加工

7、误差和确定适当的控制措施。从历史上看，质量控制的方法是基于判断检查。判断检查在年前一段时间，工业品的需求急剧增加，零件变化是常见的。熟练的工匠进行对各部分最后的改变装配确保零件的配合和功能正常。使用“lt”的做法，工匠将检查、部分调整，直到达到一个可以接受的程度。The quality control used by craftsmen was based on judgment inspection; the worker decided whether each part in an assembly was acceptable or unacceptable. If the part

8、was unacceptable, the worker would try to rework the parts or else discard them if they could not be adjusted to function properly. The relationships between judgment inspections, the production process, and associated decisions (Fig. 1 ) show that the inspection occurs after the product is made. Th

9、us, judgment inspections can detect nonconformities only after they have been generated.质量控制被木匠使用是基于判断检验；工人决定组件中的每一部分是可接受或不可接受的。如果部分是不可接受的，工人要返工零件或者他们认为如果他们不能调整到适当的功能就抛弃。判断检查之间的关系，生产过程，以及相关的决定如（图1）表示，因此，判断检查后才产生了检测不合格。今天We still use judgment inspections today. For example, a chemist may mix chemical

10、s in a test tube and hold the test tube up to a light to determine whether the chemicals are adequately mixed. According to the appearance of the mixture, the chemist may determine that mixing is adequate or that more stirring or agitation is necessary. If the mixture is of the wrong color, it may b

11、e discarded because the improper material quantities have been used.我们仍然使用判断检查。例如，一个化学家可以混合化学品的试管内用试管颜色来确定化学物质充分混合。根据混合物的外观，化学家可以确定是适当的或更多的搅拌混合或搅拌是必要的。如果混合的颜色不对，可能是因为已使用不适当的材料数量而丢弃量规检验：细化判断检查虽然Eli Whitney开发可互换的零件概念，亨利福特对其广泛的接受（Womack et al.。福特认为，部分到部分的变异性是实现真正的可互换零件的主要障碍。福特继续使用判断检查以控制部分的特点，但与计量检验达到更

12、大的一致性取代目视检查。作为一个过程来克服测量从一个位置到另一个差异，有整个福特生产系统是可追溯的标准。这有效地实施国际测量标准增加了在发展生产起关键作用。SQC inspections.统计质量控制检查The next major breakthrough in quality control was made by a team led by Shewhart at Bell Laboratories (3). This team demonstrated that variation on the production floor could be described statistic

13、ally and that statistical data could identify when a process was drifting out of control. Statistical data were useful in guiding adjustment of the process to reduce the probability that nonconforming product would be produced.在质量控制中一个重大突破是由贝尔实验室的一个小组做了休哈特控制LED（3）。该团队证明在生产变异可统计描述和统计数据确定过程中的失控漂移。统计数据

14、是有用的过程导向调整以减少不合格产品会产生的概率。The SQC inspection process illustrated in Fig. 2 had many advantages relative to judgment inspections. SQC relies on sampling inspection, which substantially reduces the amount of inspection activity. Unlike judgment inspection, SQC provides a feedback for the production pro

15、cess that helps to reduce the likelihood of nonconformities. This resulted in dramatically lower defect rates on the production floor, often in the range of 2 000 to 20 000 nonconforming parts per million (ppm). The improvement in quality achieved through SQC reduced rework, scrap, and wasted resour

16、ces.质量检验流程图2所示优于许多判断检查。SQC依赖于抽样检验，大大降低了检验活动量。不像判断检验，统计质量控制提供了一种生产的过程，有助于减少不合格的似然反馈。这导致显着降低生产车间的缺陷率，往往在2 000到20000（ppm）不合格在一亿产品中。通过SQC减少返工，报废的改善，资源浪费，提升质量。Deming estimated that 94% of the process variation stemmed from common causes and that 6% could be traced to special causes (4). These impressions

17、 were formed in a time frame when tolerancing, gaging, and SQC methods were evolving.戴明估计94%过程变化的源于常见的原因，6%可以追溯到特殊的原因（4）。这些现象是在一个时间帧时，公差，形成的测量，和SQC方法的发展。The development of SQC led to a revolutionary reduction in nonconformity rates and hence many advances and developments in the understanding of qua

18、lity and quality control. Subsequent changes, however, have generally led to incremental reductions in nonconformities. In the mid-1980s, the Department of Defense strongly stressed implementation of total quality management, a quality approach emphasizing operation “in a framework of statistical co

19、ntrol” (5). However, 60% of the companies implementing total quality management see 60%实施全面质量管理的企业有10%不合格减少（6Motorolas Six Sigma.摩托罗拉的六西格玛近期发展Perhaps the most widely recognized recent development in SQC methods is Motorolas Six Sigma (7). According to traditional SQC, a process is in control if the

20、variation measured in standard deviations ( or sigma) is less than one-third the difference between the control limits and the process mean; i.e., the traditional distance between the mean and any control limit is at least 3. However, Motorola concluded that quality leaders achieved fewer defects by

21、 reducing process variations. They estimated that defects would be cut 1000-fold if the process variation could be held to one-sixth the difference between process mean and control limits, or 6 variation control.最广泛认可的SQC方法是摩托罗拉公司六西格玛（7）。根据传统的统计过程控制，如果以标准偏差衡量的变化（或）不超过三分之一的控制范围和过程为均值之间的差异；即，意味着任何控制极限

22、之间的传统的距离至少为3。然而，摩托罗拉认为质量减少过程变异取得更少的缺陷。他们估计，缺陷将降低了1000倍，如果过程的变化，六分之一差异的过程均值和控制限，或6变化是可控的。Roughly 6 years after aggressively implementing Six Sigma methods, Motorolas average defect rate was 1000 ppm (8). Although this represented a substantial improvement in quality control for Motorola, it falls far

23、 short of the theoretical results predicted by the Six Sigma theory.大约6年后，积极实施六西格玛方法，摩托罗拉的平均缺陷率1000 ppm（8）。虽然这代表摩托罗拉在质量控制中的一个重大改进，但是它远低于通过六西格玛的理论预测理论结果。质量控制的新范式在日本，公司设定质量标准要比依靠SQC方法缺陷率较低。为了实现了类似的结果，我们必须了解他们的质量控制方法与传统的方法的不同，以及为什么这些差别是很重要的。mistakesa dominant source of defects错误-缺陷的主要原因If we study a si

24、mple process such as mixing specified quantities of two liquids in a beaker, we can describe the quantity of each of the liquids in terms of variation, similar to the traditional statistical approach, but sometimes a person may forget to put one of the liquids into the beaker, may add two measures o

25、f the same liquid into the mixture, may inadvertently forget to stir the mixture, or may select an improper chemical for the mixture. In contrast to variation, we define such mistakes or blunders as: (a) the execution of a prohibited action; (b) the failure to perform a required action; or (c) misin

26、terpretation of information essential for the correct execution of a task.我们研究了一个简单的过程，指定的数量两种液体在混合烧杯中，我们可以描述随着各种液体的数量的变化的变化，类似于传统的统计方法，但有时人会忘了把液体倒入烧杯中。添加两种相同的液体在混合物中，可能会在无意中忘了将混合料搅拌，或者选择不当的化学混合物。在对比变化后，我们定义了这样的错误或失误：（a）禁止的行为的执行；（b）未能完成规定动作；（c）执行的任务正确的信息的曲解。In many cases, mistakes are detected and c

27、orrected, often after the mistake has resulted in a nonconforming product. However, about once in 10 000 to once in every 100 000 repetitions, each type of mistake will be undetected and uncorrected (9).在许多情况下，往往造成不合格产品后，错误检测和纠正错误。，10000一个到在100 000中一个重复，每个类型的错误将被发现和未校正（9）。Types of mistakes.错误类型Altho

28、ugh each type of mistake is exceptionally rare, many types can occur. One book on mistake-proofing for manufacturing listed the following types of mistakes in order of their frequency of occurrence (1): (a) omitted processes; (b) processing mistakes; (c) mistakes setting up workpiece; (d) missing pa

29、rts; (e) wrong parts; (f) processing the wrong workpiece; (g) misoperation; (h) adjustment mistake; (i) equipment not set up properly; and (j) tools and jigs improperly prepared. These mistakes are leading causes of nonconformities. These same mistakes can occur in the clinical laboratory environmen

30、t, although they may be described in slightly different terms. For example, an omitted label on a specimen is one type of missing part.虽然这些类型的错误是非常罕见的，许多类型可能发生。一本书关于制造业上市后的错误类型发生频率顺序（1）：（a）忽略过程；（b）错误处理；（C）错误建立工件；（d）缺少的部分；（e）错误的部分；（f）处理错误的工件；（g）误操作；（h）调整的错误；（我）设备设置不正确；及（J）准备不当的工具和夹具。这些错误都是不合格的主要原因。这些

31、同样的错误，可以在实验室环境中出现的，尽管他们可能在稍有不同的描述。例如，一个略标签的标本是一种缺失的部分。工艺As the control of variation has improved, the relative importance of mistakes in the total nonconformity rates has increased. Several researchers have reached this same conclusion over the last 30 or 40 years (see review (8). For example, in th

32、e study of 23 000 production defects, Rook (10) concluded that 82% originated from mistakes.变化的控制有所改善，在总的不合格率中相错误重要性增加了。一些研究人员已经得到了同样的结论在过去的30年或40年（见评论（8）。例如，在23 000产品缺陷的研究中，认为82%（10）起源于错误。We can frequently find evidence that mistakes are dominant quality problems. For example, a 1995 article in the Journal of the American Medical Association 11 stated that 6.5% of the patients entering hospitals exp