spc和msa教程.ppt

1、SPC,Statistic Process Control,Statistic technique,Process Control,SPC,由Product Spec 到Process Parameter,由Product Spec 到Process Parameter,1.,Basic concept of SPC,Variation Control (Stability)-Statistical stable status-Process performance can be predictable Capability Over-adjustment,SPC - 品管七大手法,Histo

2、gram Check Sheet Pareto Cause and effect diagram Defect concentration diagram Scatter diagram Control chart,Basic concept of SPC,檢定是否為常態(tài)分配 管制圖的中心線 X-bar 管制圖的管制界限(信賴水準(zhǔn),信賴區(qū)間) Control Limits =信賴區(qū)間 1.集中趨勢(shì)（通常以X bar 作代表） 2.離中趨勢(shì)（通常以 作代表） 3.信賴水準(zhǔn)-被含蓋在特定範(fàn)圍內(nèi)的機(jī)率,信賴水準(zhǔn)-被含蓋在特定範(fàn)圍內(nèi)的機(jī)率,X bar,X bar,信賴水準(zhǔn),信賴區(qū)間,信賴水準(zhǔn) 信賴區(qū)間

3、 90% 1.645 95% 1.96 99% 2.575 99.73% 3 99.94% 3.5 99.99% 3.9 99.994% 3.99 ,信賴水準(zhǔn),信賴區(qū)間,P, Ppk/Cpk,Z,Basic concept of SPC,用管制圖判別special cause的準(zhǔn)則 超過管制界限 連續(xù)七點(diǎn)向上 連續(xù)七點(diǎn)向下 連續(xù)七點(diǎn)向單邊 不隨機(jī)分佈,Basic concept of SPC,Statistical stable = no special cause = Under control = process performance can be predictable 當(dāng)製程達(dá)到St

4、atistical stable時(shí)才可計(jì)算Cpk,Stable Process,Unstable Process,Basic concept of SPC,Over-adjustment : 當(dāng)對(duì)common cause所造成的變異都採(cǎi)取措施調(diào)整製程時(shí), 等於將一個(gè)special cause加入己穩(wěn)定的製程。使原本穩(wěn)定的製程變成不穩(wěn)定。,Basic concept of SPC,Ppk Ca Cp Cpk,Basic concept of SPC,Ca : Sensitive to Specification target Cp : Sensitive to variation(spread)

5、 Cpk : Sensitive to both, capability index for stable process Only common cause variation included Ppk : Performance index Both common cause and special cause variation included Cp Cpk,Basic concept of SPC,Variation : 製程輸出不可避免的差異 Inherent Variation (固有變異):由common cause所造成的變異 = R bar / d2 Total Varia

6、tion :由common cause and special cause所造成的變異 Ppk Cpk,Cpm,管制圖種類包括,計(jì)量值管制圖 計(jì)數(shù)值管制圖,Data的種類,計(jì)量值(location & spread) 計(jì)數(shù)值,計(jì)量值管制圖,X-Bar R Chart X-median R Chart X-Bar S Chart X-Rm Chart,計(jì)數(shù)值管制圖種類包括,P Pn C U,管制圖的目的,預(yù)防不良品產(chǎn)出 判定製程是否有非機(jī)遇因子造成的變異 去除非機(jī)遇因子造成的變異 減小製程的變異 改善製程 判定製程是否穩(wěn)定 製程能力分析 100%檢驗(yàn)還需耍做管制圖嗎? 如何做?,特殊應(yīng)用管制圖,

7、CUSUM cumulative-Sum Control Chart)累積和管制圖 EWMA exponentially Weighted Moving Average Control Chart)指數(shù)加權(quán)移勳平均管制圖 EVOP (Evolutionary Operation)-Part Average Test漸進(jìn)操作法管制圖 Acceptance Control Chart 允收管制圖 Modified Control Limits Control Chart修正界限管制圖 Short Run 短製程管制圖,Short Run 短製程管制圖,Guard Band Limit-GRR,Gu

8、ard Band Limit-Cpk,MSA,Measurement System Analysis,Statistic technique,MS Control,MSA,MSA,GRR v. s. CP,Accuracy & Precision,MSA,計(jì)量值 Location - Bias, Linearity, Stability Spread Repeatability, Reproducibility(GRR) Range Method Average Range Method Average Range Method (Within Part Variation) ANOVA DO

9、E 計(jì)數(shù)值-Short Method -Long Method,Bias,Stability,Linearity,Repeatability,Reproducibility,GRR-Average-Range Method,%GRR = % R&R = 100(R&R/TV) TV2 = R&R2 + PV2 R&R2 = EV2 + AV2,TV2,EV2,AV2,PV2,GRR-Range Method,GRR-Average-Range Method,ANOVA Method,MSA timing & application,MSA / APQP/ Product Realization

10、 MSA / FMEA,MSA的目的,何種情況可不必做GRR ? Cp is good, not use for process measurement, C/I not required? 只做GRR, 不做Bias, Linearity, Stability可以嗎 ? 汽車基地 ,MSA for measurement cant be repeated,The inability to test the same unit repeatedly is a major obstacle to obtaining an estimate of pure measurement error. D

11、estructive testing Measuring a temperature, viscosity Samples obtained from the product stream should be as closed together as possible With film, one might prefer adjacent samples collected in the machine direction rather than from different positions across the web Obtain pairs of samples that are

12、 as homogeneous as possible within each pair, each pair may be tested and the difference in the paired readings taken to obtain an estimate of the variation of the measurement process,Data Categories & Resolution,Data Categories & Resolution,Data Categories & Resolution,Data Categories & Resolution,

13、Data Categories & Resolution,Data Categories & Resolution,Data Categories & Resolution,Data Categories & Resolution,PT Ratio,Precision Tolerance Ratio The ratio of parameter tolerance to measurement system repeatability PT=Tolerance /EV,MSA Result,Gage resolution to spec tolerance Gage resolution to Process variation Bias Stability Linearity R&R,When should