1、基本的分析工具簡介,By Turing Xu,說明 本課程是Lean Production的課件之一 系統(tǒng)介紹了在實際工作中經(jīng)常用到的分析工具和方法 適應(yīng)對象和范圍 所有工程師, 生產(chǎn)人員, QC人員, 銷售人員, 工程人員 生產(chǎn)線班組長 課時 30分鐘 要求 能基本熟練應(yīng)用這些工具來做課余作業(yè)題,Process Map（流程分析）,為什么要作流程分析 對所要完成的事項提出清晰的理解 (改善前水準) 闡明正在發(fā)生的和將要發(fā)生的 獲取關(guān)鍵的結(jié)構(gòu)上的認知 促進對問題所在區(qū)域的識別 激勵對業(yè)務(wù)流程再設(shè)計BPR的思考 促進識別需收集數(shù)據(jù)的位置 (過程量測),流程分析是什么 流程分析是一個適合于描述一個

2、過程的作業(yè)步驟或活動的圖形化工具. 它是鑒別顧客的的一個基本計劃性工具.流程分析也稱為流程圖. 如何作出流程分析 利用正確的人員. 包括在流程中工作的人員, 顧客和供應(yīng)商. 和選擇一個中立的協(xié)助者來保持分析的中肯. 所有的團隊成員都必須參與. 數(shù)據(jù)始終必須讓每個人都能看到. 利用大字報書寫; 在會議中應(yīng)不斷地修正圖紙. 盡量不要用投影片; 一次只允許展示流程圖的一個部分. 分配足夠的時間. 流程圖應(yīng)比預(yù)計的要長. 使用通用的流程圖符號.,Process Map,Why Provides clear understanding of the execution (baseline) Illust

3、rates what is happening versus what should be happening Captures critical organizational knowledge Facilitates identification of problem areas Stimulates ideas for business process reengineering Facilitates identifying locations for data gathering (process measurement),What A process map is a graphi

4、cal means of depicting the steps or activities which constitute a process. It is a fundamental planning tool for identifying customers. Process Maps are also called flow charts. How Use the right people. Include people who work in the process, customers and suppliers. An independent facilitator to k

5、eep independence. All members must participate. Data must be visible to all personnel all the time. Start with Post-It notes; rarely is a session completed without reworking the map. Avoid transparencies; this only allows 1 segment of the chart to be displayed at a time. Enough time needs to be allo

6、tted. Process Maps take longer than is expected. Use the common charting symbols.,腦力風暴法,為什么要用腦力風暴法 腦力風暴法可以在短時間內(nèi)產(chǎn)生許多的想法/對策. 促進創(chuàng)造性思維的過程. 提出想法而不是對想法評價.,腦力風暴法是什么 一種在分析過程中能相互激發(fā)的產(chǎn)生自由的想法/對策的工具 如何進行腦力風暴法 回顧主題 “為什么, 怎樣作, 或是什么”的問題. 給每人一到兩分鐘的思考時間 要求每個人大聲說出他們的的想法 (重要的是: 在這個過程中不對想法作任何討論). 記下每一個想法. 合并相似的想法并且對完成的列

7、表進行討論. 利用一定的方法綜合多數(shù)意見來進行排序篩選.,Brainstorming,Why Brainstorming produces many ideas/ solutions in a short time. Facilitates the creative thinking process. Separates idea generation from the judging of the ideas.,What A freeform method of generating unconstrained ideas/solutions and equalizing involveme

8、nt in the analysis process How Review the topic why, how, or what questions. Give everyone a minute or two of silence to think about the question. Invite everyone to call out their ideas (important: no discussion of ideas until session is complete). Write down every idea. Consolidate like ideas and

9、discuss complete list. Use consensus building tools to assist in prioritization.,名稱歸類法,為何要進行名稱歸類法 NGT（名稱歸類法）可以在短時間內(nèi)產(chǎn)生許多的想法/對策 組織對問題的思考, 而不是人; 開通多方的溝通渠道; 允許不同意見 能夠達成一致多數(shù)同意的最終結(jié)論. 特別是有重大爭議的問題,名稱歸類法 一個類同于腦力風暴的方法.一種組織收集想法的辦法, 調(diào)查一個較小的團隊的觀點, 及對這些想法進行篩選排序. 怎樣進行名稱歸類法 介紹現(xiàn)有問題. 產(chǎn)生想法, 安靜地思考5 到 10 分鐘, 不要討論. 匯集想法,

10、 輪流地, 每次一個想法, 寫在貼紙上并張貼. 處理/層別這些想法 消除重復(fù)的觀點, 合并相近的觀點. 不要過多的為觀點的解釋和協(xié)商進行討論. 專注于對意見的分類, 不要爭執(zhí). 5個5個地清點意見數(shù)量. 準備相應(yīng)數(shù)量 X 個卡片其中 X =上一步識別出來的意見數(shù)量. 在卡片的右下角按照由高到低的優(yōu)先順序?qū)懮蠌?到 X的編號. 按照卡片上的編號填上選出第一優(yōu)先的項目編號.再選出第二優(yōu)先的項目, 依次類推. 直到所有的卡片填滿. 列出分數(shù)表并用一張柏拉圖表示,Nominal Group Technique,Why NGT produces many ideas/solutions in a sho

11、rt time Structured to focus on problems, not people; to open lines of communication; tolerate conflicting ideas Builds consensus and commitment to the final result. Especially good for highly controversial issues,What A technique similar to brainstorming. A structured approach to generate ideas, sur

12、vey the opinions of a small group, and prioritize those ideas. How Present issue, instructions. Generate ideas, 5 to 10 minutes of quiet time, no discussion. Gather ideas, round-robin, one idea at a time, written on flipchart and posted. Process/clarify ideas duplicates are eliminated, like ideas ar

13、e combined. Limit discussion to brief explanations and brief agreement statements. Focus on clarification of meaning, not arguing points. Count each idea and divide by 5. Prepare X number of cards where X = the number identified above. Write a number from 1 to X on the bottom right corner of each ca

14、rd with the highest number being the highest priority. Choose the first priority and put the item number on the paper with the highest card number. Pick the second priority, etc. until all cards are filled out. Tabulate scores and present on a Pareto chart,柏拉圖,為什么要做柏拉圖 此方法基于柏拉圖原理 (80/20), 即我們問題的80%是

15、可以用 20% 的原因來解釋. 柏拉圖能辨別首先應(yīng)該處理的“少數(shù)但重大”的問題(或原因) .,什么是柏拉圖 一種逐項遞減的柱形圖, 最大的一項放在最左邊. 每個柱子代表一種問題（或原因）. 顯示了每種原因?qū)φ麄€問題的影響程度. 怎樣做柏拉圖 列出所有的相關(guān)因素. 用同樣的測量單位對這些因素進行量測. 按照這些因素的量測值排序, 而不是它們的等級. 計算這些項目的累積占有率及其測量值來作出柱形和連線圖. 首先解決最重要的問題. 如果可以獲得數(shù)據(jù)的話，可以從不同分類的角度進行反復(fù)多次的柏拉圖分析.,Pareto Chart,Why This technique is based on the Pa

16、reto principle (80/20), which states that 80% of our problem can be explained by just 20% of the causes. The Pareto chart makes clear which vital few problems (causes) should be addressed first.,What A bar chart in which the bars are arranged in descending order, with the largest to the left. Each b

17、ar represents a problem (cause). The chart displays the relative contribution of each cause to the total problem. How List all elements of interest. Measure the elements, using the same unit of measurement for each element. Order the elements according to their measure, not their classification. Cre

18、ate a cumulative distribution for the number of items and elements measured and make a bar and line graph. Work on the most important elements first. Continue for several iterations if data is available or can be measured.,因果關(guān)系圖 (魚骨圖),為什么要作魚骨圖 當問題很難用定量的數(shù)據(jù)分析時則用魚骨圖找出并層別影響一個過程的因素. 通過討論協(xié)助找到根原因 根據(jù)問題的癥狀確定

19、真正的原因. 統(tǒng)一一個團隊對問題理解的不同層次.,什么是魚骨圖 用于表示潛在的原因?qū)ζ浣Y(jié)果（問題）的影響關(guān)系. 怎樣作魚骨圖 定義問題. 確定原因的大分類. 大分類可能包括: 數(shù)據(jù)和信息系統(tǒng), 價值, 環(huán)境, 硬件, 材料, 量測, 方法, 人員, 培訓等. 用腦力風暴法分析導致大分類問題的詳細原因. 對每個大分類至少要連問5次“為什么”. 刪除不適用的原因. 對這些原因進行討論并確定最重要的原因. 首先解決最重要的原因. 如果必要的話, 對根原因進行重復(fù)操作.,Cause and Effect Diagram (Fishbone),Why Sort and relate the factor

20、s affecting a process while little quantifiable data is available. Discussion guide to assist in determining root causes. To determine the real cause to the problem versus a symptom. To identify a teams level of understanding.,What Represents the relationship between an effect (problem) and its pote

21、ntial causes. How Name the problem. Decide the major categories for causes. Major causes may include: data and information systems, dollars, environment, hardware, materials, measurements, method, people, training, etc. Brainstorm for more detailed causes. Ask why each major cause happens at least 5

22、 times. Eliminate causes that do not apply. Discuss the causes and decide which are most important. Work on most important root causes. Perform another iteration for root causes, if necessary.,投票表決法,為什么要用投票表決法 發(fā)揮團隊的作用或者通過不受其他人影響的投票讓每個成員自由表達對問題輕重程度的看法. 減少列表中的被選擇項 (to得出最重要的一到兩項).,什么是投票表決法 一種用于篩選或限定一組被

23、選擇項的方法. 可以用來作列表排序. 怎樣作投票表決法 列出并清點所有被選擇項 (例如, 對被選擇項編號). 發(fā)給每個成員3個投票權(quán) (或任何相對被選擇項數(shù)量合理的數(shù)量). 按照以下規(guī)則每個成員進行投票: 每個成員可以對任何選擇項投全部的, 或部分的選票甚至不投票. 當所有的票投完, 該成員就算完成. 按照得票數(shù)的順序排列選擇項的順序. 重復(fù)上述過程知道達到期望的結(jié)果.,Multivoting,Why Improve the teams ability or allow each member to express strength of opinion through voting with

24、out risking judgment by group. To narrow down a list of alternatives (to the one or two most important).,What A method for selecting or narrowing a set of alternatives. Can be used to prioritize a list. How List all alternatives and count them (i.e., N# of alternatives). Give each member N/3 votes (

25、or whatever # seems reasonable given the number of alternatives). Have each member cast his / her votes given these guidelines: The member can cast all, some, or none of his votes for any alternative. When all votes (the sum) are cast, the member is done. Rank order the alternatives based on the num

26、ber of votes received. Repeat the process until desired result is accomplished.,查檢表,為什么要用查檢表 通過數(shù)據(jù)呈現(xiàn)使答案更加明顯. 幫助分析數(shù)據(jù)以確定問題的真正答案. 從數(shù)據(jù)中獲得信息. 收集與問題相關(guān)的數(shù)據(jù)和事實.,什么是查檢表 一種簡單的容易說明結(jié)果的記錄表. 不僅能記錄數(shù)據(jù)而且能同時分析數(shù)據(jù)的趨勢. 怎樣作查檢表 對項目所需相關(guān)的特殊信息提出質(zhì)詢. 運用適當?shù)臄?shù)據(jù)分析工具 (例如, 柏拉圖, 柱狀圖). 確定包含數(shù)據(jù)的收集點. 了解數(shù)據(jù)的收集者和他們的所處的環(huán)境. 設(shè)計數(shù)據(jù)收集的表格 (盡量簡化). 培訓

27、數(shù)據(jù)收集者 包括研究的目的 審查數(shù)據(jù)收集的過程.,Checksheet,Why To present data in a way that clearly communicates the answer. To help analyze the data to determine the factual answers to the question. To generate information from data. To collect the data and facts relating to the question.,What A simple recording form th

28、at has been designed to readily interpret results from the form itself. A well-designed form does more than simply provide a record of the data it allows simultaneous analysis trends in the data. How Formulate good questions that relate to specific information needs of the project. Use appropriate d

29、ata analysis tools (e.g., Pareto, histogram). Define comprehensive data collection points. Understand data collectors and their environment. Design data collection form (keep it simple). Train data collectors including purpose of study Audit the collection process.,推移圖,為什么要用推移圖 數(shù)據(jù)可視化. 監(jiān)控一個過程看是否長期的平均

30、值正在有所變化. 關(guān)注過程中的重要的真實的變化. 識別平均值中有代表意義的趨勢或變動.,什么是推移圖 在特定的一段時間內(nèi)根據(jù)觀察點的變化簡單地顯示其可能的趨勢. 怎樣作推移圖 確定被量測的特性值. 確定被測量值的時間間隔. 選擇所用的度量單位. 按照發(fā)生的順序在圖中排列標示點并用一段一段的直線將其連接起來. 分析數(shù)據(jù)查找不符隨機性的異常事件. 一般來說分布在平均值線的上下兩端的點數(shù)理應(yīng)大致相等.,Run Chart,Why To visually represent data. To monitor a process to see whether the long range average

31、 is changing. To focus attention on truly vital changes in the process. To identify meaningful trends or shifts in the average.,What The simplest possible display of trends within written observation points over a specified periods of time. How Identify the measurement to be made. Decide on the time

32、 interval of the measurements. Select the scale to be used. Plot points on the order in which they occur and connect with straight lines. Analyze the data looking for non random events. It is expected that we should find an equal number of points falling above and below the average.,散布圖,為什么要作散布圖 確定兩

33、個數(shù)據(jù)系列之間的關(guān)系. 例如, 是否積壓的工作量會對計算機數(shù)據(jù)輸入的失誤率有所影響. 數(shù)據(jù)間的關(guān)系可以通過一個散布圖進行非數(shù)學性質(zhì)的評估.,什么是散布圖 一種通過測繪兩種參數(shù)以確定是否其中有所關(guān)聯(lián)的方法. 這種工具有時也被稱為交叉圖或散點圖. 怎樣作散布圖 收集可能會有關(guān)聯(lián)的數(shù)據(jù)對. 這對數(shù)據(jù)在散布圖中用X-Y坐標系標示. 每一個軸分別用作兩種進行比較的數(shù)據(jù)系列中的一種. Y 軸通常留作表示我們想要預(yù)知的特性值, 例如, 失誤率. X 軸用于表示影響特性值變化的自變量, 例如., 工作積壓量.,Scatter Diagram,Why To determine relationships bet

34、ween two sets of data. For example, whether work backlog affects the error rate of computer data entry. Data relationships can be non-mathematically evaluated by using a scatter diagram.,What A graphical method of plotting two parameters to determine if there is a relationship between the two. This

35、tool is sometimes called a cross plot or scatter plot. How Collect data that are suspected to be related. The data pairs are plotted on X-Y axis as a scatter diagram. Each axis is used for one of the two sets of data being compared. The Y axis is usually reserved for the characteristic we would like

36、 to predict, e.g., the error rate. The X axis is used for the variable which makes the prediction, e.g., work backlog.,直方圖,為什么要作直方圖 聚焦于一個問題的原因. 數(shù)據(jù)總是可以顯示變異. 變異表現(xiàn)的形式稱為分布. 變異由過程中的因素的差異造成. 不同的現(xiàn)象將呈現(xiàn)不同的模式.,什么是直方圖 一種描述一系列數(shù)據(jù)離散性的圖形工具. 各個柱形自然的排布可以輕易地看出在一個簡單的數(shù)據(jù)表中所不能看出的圖形趨勢. 怎樣作直方圖 找出數(shù)據(jù)最大值與最小值之間的差距. 將其劃分為 7 到 1

37、4個可能的區(qū)間. 設(shè)定區(qū)間的寬度使樣本中的最小量測值包含其內(nèi). 記下包含全部數(shù)據(jù)的所有區(qū)間的界限. 每觀察到一個數(shù)據(jù)即將其標記在所對應(yīng)的區(qū)間內(nèi)從而形成一個發(fā)生頻率的分布.,Histogram,Why Sheds light on the cause of a problem. Data almost always shows variation. Variation displays patterns called distributions. Variation is due to differences in process factors. Different phenomena wil

38、l have different patterns.,What A graphic summary of variation in a set of data. The pictorial nature of the histogram enables one to see patterns that are difficult to see in a simple table of numbers. How Find the difference between the min and max data. Divide by 7 and 14 to indicate the possible

39、 intervals. Set up an interval of width chosen which will include the min measurement in the sample. Write limits for all intervals to include all data. Enter a tally mark for each observation in the interval thus forming a frequency distribution.,管制圖,為什么要作管制圖 監(jiān)控和控制過程. 很輕易地可以判斷過程是否受控l. 可以知道何時需要調(diào)整過程.

40、 可以辨別正常變動還是異常變動. 可以確定過程能力. 可以從變動的圖形中發(fā)生的概率預(yù)測到異常點的發(fā)生.,什么是管制圖 一種測量并分析過程變動情況的統(tǒng)計應(yīng)用技巧. 這種統(tǒng)計過程控制（SPC）的圖形診斷可以分析問題和改善過程. 怎樣作管制圖 確定管制圖的目的. 選擇圖表所用的變量或特征值. 選擇子組大小及數(shù)據(jù)的采集頻率. 計算平均值, 標準差, 管制上下限及畫圖. 從圖中獲得初步的結(jié)論. 指出哪些受控或失控.過程中什么正在發(fā)生以及什么將要發(fā)生它們之間的聯(lián)系. 根據(jù)圖中的啟示采取相應(yīng)的行動.,Control Chart,Why Monitor and control a process. At a

41、glance, can determine whether process is in control. To know when to adjust a process. To know the difference between assignable and chance variation. To determine process capability. To predict from the pattern of variability the percentage of occurrences that fall outside expected values.,What The

42、 application of statistical techniques for measuring and analyzing the variation of a process. The diagnosis of the SPC graphs allow one to analyze problems and improve the process. How Decisions on the objectives of the charts. Choice of variable or attribute charts. Select subgroup size and freque

43、ncy of data collection. Calculate average, standard deviation, control limits and plot the charts. Draw preliminary conclusions from the charts. Indication of control or lack of control. Relationship between what process is doing and what its supposed to do. Take action suggested by the charts.,親和圖,

44、為什么要作親和圖 當需要組織突破問題的困擾時. 有助于突出中心主題. 當問題的表象十分雜亂時. 當需要突破傳統(tǒng)的思路時.,什么是親和圖 一種將事實, 觀點和問題點組織成自然的組合來協(xié)助分析一個復(fù)雜的問題. 將所有意見記載于卡片上然后重新排列直到形成有效的組合. 怎樣作親和圖 集合適當?shù)膱F隊. 清楚地描述問題的所在. 將腦力激蕩所產(chǎn)生的想法記在卡片上. 清楚地表達想法并將卡片貼在墻上. 將有關(guān)聯(lián)的卡片集中成組. 寫出卡片的標題. 完成圖表.,Affinity Diagram,Why When breakthrough is needed to help organize. To help dev

45、elop central theme. Facts on a problem are not well organized. Breakthrough in traditional thinking is needed.,What A tool for organizing facts, opinions and issues into natural grouping as an aid to diagnose a complex problem. The inputs are listed on cards which are then rearranged until useful groups are identified. How Assemble the right t