规划过程基线之程序基线描述课件

)1)2)3nThe objective is to report how the process is performing at present.nShort Term VariationnLong term process capabilitynVariationnCenteringnStabilitynHow are the independent variables changing as the process runs its course?Data collection will be designed to answer these questions)4n以汇报目前程序运行情况为目标。n短期差异n长期程序能力n差异n集中n稳定n各独立的变量是如何在程序运行过程中发生差异的。数据收集专为回答这些问题而设计。数据收集专为回答这些问题而设计。)5nThere are two types of variation.nCommon Cause VariationnVariation that is inherent in the existing processnAssignable(Special)Cause VariationnVariation that can be segregated and assigned to a specific cause.Each“Box”represents a subgroup of data100080060040020050403020100-10Long Term Variation(Assignable Cause)Between SubgroupsShort Term Variation(Common Cause)Within SubgroupsEach“Box”represents a subgroup of data100080060040020050403020100-10Long Term Variation(Assignable Cause)Between SubgroupsShort Term Variation(Common Cause)Within Subgroups100080060040020050403020100-10Long Term Variation(Assignable Cause)Between SubgroupsShort Term Variation(Common Cause)Within SubgroupsOur process analysis methods will be planned to enable us to separate and evaluate both kinds of variation.)6n差异存在两种。n常规原因的差异 n差异固定存在于现有程序中。n（特殊）原因的差异 n差异可以被隔离并被归因于一特定原因。Each“Box”represents a subgroup of data100080060040020050403020100-10Long Term Variation(Assignable Cause)Between SubgroupsShort Term Variation(Common Cause)Within SubgroupsEach“Box”represents a subgroup of data100080060040020050403020100-10Long Term Variation(Assignable Cause)Between SubgroupsShort Term Variation(Common Cause)Within Subgroups100080060040020050403020100-10Long Term Variation(Assignable Cause)Between SubgroupsShort Term Variation(Common Cause)Within Subgroups程序分析方法将使我们能有效分离及评估两种差异。程序分析方法将使我们能有效分离及评估两种差异。)7nThis variation is present in every process.nIt cannot be controlled or reduced with the existing technology.nIt limits the process capability.nIn process baselining,we measure this variation by the amount of variation in the short term subgroups,often called rational subgroups.nsST will record the common cause variation and ZST will report the Short Term Capability.nAnother term for Common Cause Variation is Residual.nSince this variation is related to process technology,a poor ZST will generally indicate a need for addressing part or process design issues.)8n此差异存在于每一程序中。n运用现有技术是不能控制或减少此差异。n此差异限制了程序能力。n在程序基线内，我们运用短期子群内的差异量来测量差异，短期子群即为通常所称的合理子群。nsST记录常规差异，ZST报告短期能力。n常规原因的差异的另一术语为残差。n由于差异与程序技术相关联，因此一个相对差的ZST会显示需要重新描述部分程序设计。)9nThis variation is the result of outside influences on a process that cause it to shift and drift.nIt is potentially controllable with the existing technologynIt will affect the process performance over time.nWe plan the process baselining in such a way as to maximize the opportunity for capturing the affect of assignable cause variation.nsLT will record the total assignable cause variation.ZLT will report the total long term capability.nIn order to address this variation,well need to know how the various vital factors contribute to causing it.This will be the objective of Analyze.)10n这种差异是外来因素影响程序运行的结果。n运用现有的技术有潜在的可能来控制此差异。n随着时间的推移，此差异将会影响到程序运行。n我们所计划的程序基线是以最能捕获可归属原因的差异为基本。nsLT记录总体可归属原因的差异，ZLT能报告总体长期能力。n为了描述此差异，我们需要知道各种对于引起此差异起相当作用的重要因素。这就是分析的目的。)11nTransportationnType vehicleType RoadsnLocation of Terminus WeathernRaw Materials/SuppliesnPercent RegrindMaterial PropertiesnPeoplenExperienceTraining nEnvironmentnTemperatureHumiditynManufacturingnTool WearShut heightnCommercialnInternet connection speedKeystrokes)12n运输n车辆类型道路类型n终点位置 天气n原料/供应n再研磨百分率原料属性n人n经验培训 n环境n温度 湿度n制造n工具磨损模具冲压尺寸n商用n互联网连接速度n键击)13nA Rational subgroup is a small collection of samples taken over a short period of time or in similar conditions.We believe the data is as free as possible from multiple levels of assignable causes.nIt is a snapshot of the process.nMultiple Rational Subgroup samples are the basis for quantifying process variation to reflect both short term and long term capability.nCollecting data in this manner allows us to characterize:nTechnology(Short Term)vs Control(Long Term)nCentering vs.SpreadCharacterizing the issues is a Giant Step towards resolving them)14n合理子集指的是在短时期或在相似条件下获得的一小组样本的集合。我们认为这些数据可以自由获取于不同级别可归属的原因。n这是对程序的快照。n多重合理子集样本是量化程序差异的基本，可同时反映短期和长期的能力。n这样的数据收集方式使我们能够描述：n技术（短期）vs 控制（长期）n集中 vs.离散描述问题是解决问题的重大步骤描述问题是解决问题的重大步骤)15nZST(Z Short Term)is a measure of technological capability.nInternal process variation drives short term capability.nZST is calculated assuming the process is centered on the process target.nThe goal for ZST is 6.0.nShort Term Capability is also known as nProcess EntitlementnProcess PotentialnTechnology of the ProcessThe measure ZST is the best your process can be.)16nZST(Z 短期）是技术能力的测量方法。n内部程序变化驱动短期能力。nZST的计算基于下列假设，即该程序是集中以程序目标为中心。nZST的目标值为6.0。n短期能力应被概括为n程序应有能力n程序潜能n程序技术ZST能使你的程序最优化能使你的程序最优化)17nZLT(Z Long Term)is a measure of the total variation of the process output over time.nData collection period must be long enough or under a range of conditions that allow opportunity for external factors to influence the process.nBoth Process Variation and Process Control drive long term capability.nZLT is calculated using the actual measured mean of the data.nThe Long Term Capability represents what the customer sees.nThe goal for ZLT is 4.5nZLT is also called Process Performance)18nZLT(Z 长期)是程序结果总体差异的测量方法。n数据收集时期必须足够长，或在一定范围条件下允许外来因素有机会影响到程序发展。n程序差异和程序控制驱动长期能力。nZLT 使用实际测量获得数据的平均做计算依据。n长期能力代表客户所见到的部分。nZLT 的目标值是 4.5。nZLT 也被叫做流程绩效。)19Z Short TermZ Short Term is measured using observed variation and the target mean.The Z reported is due to short term variation only.Z Long TermZ Long Term is measured using both the observed variation and the observed mean.Because the distance between the mean and spec limit may be different for each spec limit,Z Long Term is calculated for each spec limit.STSTTargetTSLZ)(LTLTSLZZ Bench is the Z Score of the summation of the probabilities of a defect at both the upper and lower spec limit.)(totalBenchdPZZLSLUSLtotaldPdPdP)()()(where)20Z 短期短期Z 短期是运用观测所获差异值和短期是运用观测所获差异值和目标平均值来测算。目标平均值来测算。所获所获Z仅归因于短期差异。仅归因于短期差异。Z 长期长期Z 长期是运用观测所获差异值和长期是运用观测所获差异值和观测所获平均值来测算的。观测所获平均值来测算的。由于对于每个规格限制，平均数由于对于每个规格限制，平均数和规格限制的间隔有可能是和规格限制的间隔有可能是不同的，因此不同的，因此Z长期需根据每长期需根据每个规格限制分别计算。个规格限制分别计算。STSTTargetTSLZ)(LTLTSLZZ Bench是缺陷在高规格限制和低规格限制概率总合的Z Score。)(totalBenchdPZZLSLUSLtotaldPdPdP)()()(where)21nThe measure of the ability to control the process is actually the difference between the Long Term Capability and the Short Term Capability.LTSTSHIFTZZZOn average ZSHIFT is about 1.5 for a typical process)22n控制程序能力的量度事实上即是长期能力和短期能力的差。LTSTSHIFTZZZ平均起来，平均起来，一个典型程序的一个典型程序的ZSHIFT大约为大约为1.5。)23nHow am I doing?nCompare the calculated ZSHIFT to 1.5.nIf less than 1.5,your process control is better than average.nIf significantly greater than 1.5,the control is worse than average.nIf only long term data is available assume the shift and add 1.5 to ZLT to estimate ZST.nTypical Historical Data situationnIf only short term data is available assume the shift and subtract 1.5 from ZST to estimate ZLT.nTypical new tooling qualification situation5.1LTSTZZ5.1STLTZZnHow am I doing?nCompare the calculated ZSHIFT to 1.5.nIf less than 1.5,your process control is better than average.nIf significantly greater than 1.5,the control is worse than average.nIf only long term data is available assume the shift and add 1.5 to ZLT to estimate ZST.nTypical Historical Data situationnIf only short term data is available assume the shift and subtract 1.5 from ZST to estimate ZLT.nTypical new tooling qualification situation1.5 ZSHIFT is an approximation.Use it with caution!)245.1STLTZZn我做得怎么样?n将计算而得的ZSHIFT数值与 1.5做比较.n如果小于1.5,则你的程序控制优于平均水平。n如果较1.5大许多，则程序控制次于平均水平。n如果仅有长期数据可用，则可假设shift并在ZLT值上加1.5来估计ZST值。n典型历史数据情况n如果仅有短期数据可用，假设shift并在ZST值上减去1.5来估计ZLT值。n典型的新工具资质条件1.5 ZSHIFT 是一近似值。请谨慎使用！是一近似值。请谨慎使用！5.1LTSTZZ)25nLong Term CapabilitynActual Process PerformancenZLTnCPKn6s meansnZLT=4.5CPK=1.5nShort Term CapabilitynThe Best the process can benZSTnCPn6s meansnZST=6.0nCPK=2.0)26n长期能力n现实程序绩效nZLTnCPKn6s 方式nZLT=4.5nCPK=1.5n短期能力n程序状态最优nZSTnCPn6s 方式nZST=6.0nCPK=2.0)27nWith the project advocacy team,identify the independent(X)variables to be recorded as you collect baseline data.nEvaluate measurement capability for the identified Ys and Xs.nDesign the plan to provide opportunity for the identified independent variables to change as much as possible.nTake manufacturing process subgroups over different shifts,after different setups,with different operators or after raw materials have changed.nDesign the plan so that variation within subgroups is limited to common cause.nRecord the value for the output characteristic(Y)and the values for the applicable Xs for each observation.nThe more data you keep the better.It will improve your ability to find and identify vital factors(Xs)in the analyze phase.)28n由项目成员确立识别独立（x）变量，并记录为基线数据。n对于Ys和 Xs进行测量能力的评估。n设计能尽量提供机会，使所要识别的独立变量发生变化的有效方案。n在不同的轮班期内，装备、操作者改变后或原料更换后，选取不同制造程序子集。n设计能将子集内的差异仅限于常规原因的计划。n对于每次观测，记录结果（Y）的值和可相应的X值。n记录数据越多越好。有利于在分析阶段中，提高发现和识别重要因素（Xs）的能力。)29nCollect enough data so as to be sure the estimate of process performance has stabilized.nClarify with the team that they must not bias the results by changing the way they do things.Do not TWEAK!nIf the Project“Y”is discrete,you will likely need a significantly larger sample size.)30n收集足够数据以确保程序绩效估计的稳定性。n向项目组明确，不能因偏好某一结论而改变行事方式。n如果 Y是离散的，你很有可能需要更大的样本。)31nRational Subgroups are often best defined by using similar tools as structured problem solving.A Fishbone diagram with the five Ms and and E can be a great place to start.nManpower(People)nMethodnMeasurementnMachinerynMaterialnEnvironment)32n合理子集通常被恰当地定义为用相似工具结构化地解决问题。如下所述，即可恰如其分地说明合理子集。n人手（人）n方式n测量n机器n材料n环境)33nIn manufacturing,an appliance plant used extruded gaskets manufactured at a supplier.In order to properly evaluate the process and the part quality,it is necessary to establish a Baseline measurement using Rational Subgroups.Potential XsRational SubgroupsExtrudersCollect parts made on each available extruderCut five(5)measurable cross-sections of the part from a five foot length for a single subgroupMultiple ShiftsCapture data 2 3 times per shift for all shiftMaterial lots of rubberAllow multiple batches of raw materials,recording CTQ batch characteristicsExtruder Set upAllow for at least to process shutdowns and new setups.Allow time for process to stabilize after start up prior to taking measurements)34n在制造业，一家设备制造工厂使用在另一供应商生产的压延垫衬。为了能 恰当评估过程和部分品质，有必要使用合理子集来建立基线测量法。潜在的潜在的 Xs合理子集合理子集挤压机收集每个挤压机上生产出来的压延垫衬在压延垫衬上横切出五件可量度的样本来作一子群 多班工作制对于所有轮班，每次轮班提取2-3次数据。橡胶原料使用多批原料，记录CTQ批量特征挤压机器装配至少允许停工和重新的装配。.在机器启动后，给予时间使程序达到稳定运行才作测量。)35nIn Commercial Projects a similar process is used.A key difference is that factors are generally categorical rather than measurable.For a Billing Error Problem,the plan might look more like this.Potential XsSubgrouping PlanCustomer Record the customer for the billBilling methodCapture billing information from both electronically generated bills and paper transactionsOperator Capture bills generated by all billing department agentsProduct LineCapture bills generated for product in each of the key product lines)36n使用相似的过程在商业项目中。关键的区别在于有关因素是分类的而不是可测量的。对于一个票据错误，这个计划更有可能是如下的。潜在潜在 Xs子集计划子集计划客户根据帐单，记录客户结算方式从电子记帐系统和手工记帐系统同时收集支付信息。操作人员 收集所有产生于各单证部门的帐单。生产线收集产品帐单，该产品生产于各重要生产线上。)37You are working on a project to reduce the cycle time for delivery of product across the Border from Mexico to Texas.The Project Y is“Hours from release from Juarez plant to departure from El Paso Distribution Centers.There are two distribution Centers in El Paso.There are three International Portals from Juarez to the US.Develop a Baselining plan to collect data and characterize the current process.Brainstorm a list of potential XsIdentify which Xs you will subgroup across.Why?Determine which Xs you will just record.Why?What will be the size of the subgroups?Why?How many subgroups.Why?)38你正从事一个项目，该项目目的是减少从墨西哥到德克萨斯货物运输周期。项目Y是从Juarez工厂出货到离开El Paso 配送中心的时间。在EL Paso有两个配送中心。从Juarez到美国有三个国际入口。收集数据设计基线计划，描述当前程序。集体讨论潜在Xs清单确定哪些Xs为子群。为什么？决定需要记录哪个 Xs.为什么?子群大小如何。为什么？有几个子群。为什么