Problem Solving Tree(ENG)

1、Delphi Confidential Main Topics Kevin Liu(劉強(qiáng)劉強(qiáng)) - Shainin Trainer - MB:- Tel: 86-20-82268388 Ext.8165 - Fax: 86-20-82268715 Delphi Confidential Main Topics Problem Solving Toolbox - 5 Why - Is-Is Not Matrix - Problem Solving Tree Delphi Confidential Problem Solving Tool box - 5 Why Delp

2、hi Confidential Linkage between 5-Why at a minimum, implement temporary measures. Part 4: Problem Prevention _Error Proof the root cause. _Verify for effectiveness. _Capture Lessons Learned. Delphi Confidential 5 Why-Template Lessons Learned: Define Problem Use this path for the specific nonconforma

3、nce being investigated WHY? WHY? Use this path WHY? to investigate why the problem was not detected WHY? WHY?WHY? Use this path WHY? to investigate the systemic root cause WHY? WHY?WHY? WHY? WHY? WHY? Corrective ActionsDate Root Causes A B A B C C Reference No. _ (Spill, PR/R) Date: _ Team members:

4、Person responsible: Delphi Confidential 5 Why-Example 1 Directly links the root cause to the Problem Description Machine stops in the middle of its operation (Symptom) Fuse keeps blowing Solution: Install pump filter on all machines THEREFORE THEREFORE THEREFORE THEREFORE THEREFORE Why does machine

5、stop? Why does fuse keep blowing? Circuit over- loaded Why is the circuit overloaded? Excessive drag on bearings Why is there excessive drag? Insufficient circulating oil Why is there insufficient circulating oil? Pump is clogged w/ metal chips No pump filter Why is the pump clogged w/ Metal chips D

6、elphi Confidential 5 Why-Example 2 小甜餅的味道真差小甜餅的味道真差 問題問題 用的是鵝蛋而不是雞蛋用的是鵝蛋而不是雞蛋 說明書沒有指明用什么說明書沒有指明用什么 蛋蛋 根本根本reason 添加材料有問題添加材料有問題 小甜餅沒有做熟小甜餅沒有做熟 為什么為什么 所所 以以 Delphi Confidential 5 Why-Example 3 杰斐遜紀(jì)念館的大理石杰斐遜紀(jì)念館的大理石 由于頻繁清洗損壞嚴(yán)重由于頻繁清洗損壞嚴(yán)重 問題問題 豐富的食物資源豐富的食物資源(小昆蟲小昆蟲) 夜晚照明紀(jì)念館夜晚照明紀(jì)念館根本根本reason 豐富的食物資源豐富的食物資

7、源(蜘蛛蜘蛛) 太多的鳥糞太多的鳥糞 為什么為什么? 所以所以 Delphi Confidential 5 Why-Case Study ACTIVITY: FLEX INDUSTRIES CASE STUDY Directions: Use the Delphi 5-Why Root Cause Analysis to evaluate the Flex Industries Case Study below and on the following pages. Review the 5-Why analysis template to record your information. U

8、se only the information provided in the case study to complete the 5-Why. Do not try to re-engineer the rivets or solve the technical problems in the case study. The purpose of this activity is to use the 5-why template to organize the given data and uncover the true root cause(s). FLEX INDUSTRIES C

9、ASE STUDY Key Players Jake Quality Manager Janet CSE, Winding River Plant SamTeam Leader, East Line JudyQA Auditor, East Line BenMachine Operator Delphi Confidential 5 Why-Case Study Background Jake Ryan is the Quality Manager at Flex Industries. Flex is a component supplier that manufactures metal

10、stampings and light assembly products. The company has a reputation for supplying high quality parts on a consistent basis. Seldom has there been a customer complaint. Flex has Quality representatives called Customer Support Engineers (CSEs) at every customer assembly plant. The CSEs report any prob

11、lems to Jake for investigation and follow-up. At 7:00 a.m. this morning, Jake received a call from Janet, CSE at the Winding River Assembly Plant. Janet informed him that the customer had found five defective stabilizing brackets on second shift last night. She checked the remaining inventory and th

12、ere were no defects in the remaining 326 pieces. The manufacturing sticker on the back of the brackets indicated that they were made by the second shift operator. Normally, the stabilizing bracket is fastened to the regulator motor with three rivets. The five defective brackets had only two rivets i

13、n them. The lower set of rivets on all five brackets was missing a rivet. This could only happen if the automated rivet machine did not place rivet in position . This was the first time that the problem occurred. Delphi Confidential 5 Why-Case Study Jake set-up containment procedures at the plant wa

14、rehouse to sort for discrepant materials. As of this morning, two more defective brackets had been found in the remaining 2019 pieces of inventory at Flex. GoodBad Delphi Confidential 5 Why-Case Study Cause Investigation Jake went out to the floor to talk with the team leader of the two rivet lines

15、(East and West) and the area quality assurance auditor. He informed Sam (the team leader) of the quality problem and asked him to identify the line that ran the stabilizing bracket assembly. Sam directed Jake to the East line which ran Winding River assembly brackets only. At the East Line, he spoke

16、 with Judy (the QA Auditor for the area) and asked to see the quality log sheets. Jake and Judy reviewed the Nov. 11th log sheet and could not find anything out of the ordinary. He asked Judy to set-up in-house containment procedures to sort for any discrepant material in the finished goods area. Th

17、en Jake looked at the control plan to see what the requirements were for checking the bracket. The only required check was to do a dimensional verification at machine set up. Jake decided the next thing to do was to go to the document control center and look at the PFMEA for the bracket. He needed t

18、o find out what failure modes were called out on the PFMEA for missing rivets, thinking this might help him in his investigation. Unfortunately, the only failure mode called out on the PFMEA for missing rivets was poor set-up, and he already knew the set-up for the bracket had been approved. He even

19、 had a sample part from the set-up and the rivets looked fine. Delphi Confidential 5 Why-Case Study Next, Jake tried to locate the second shift operator whose clock number was on the defective parts. Since that operator was gone, Jake spoke with the current machine operator (Ben). He asked Ben about

20、 any recent difficulties with the rivet machine. Ben said that he hadnt noticed anything out of the ordinary. Ben also mentioned, however, that there had never been any quality bulletins posted in the two years that this particular part has been running. Jake decided to stay in the area to watch the

21、 machine run for a while. After about 15 minutes, he watched Ben dump rivets into the feeder bowl to prepare for the next run. Shortly after restarting the rivet operation, Ben walked over to another riveter and came back with a steel rod. Ben poked around the rivet chute and then continued working.

22、 Jake approached Ben and asked him about the steel rod. Ben replied that from time to time the chute gets jammed and he has to clear it out. If the chute got jammed, the automatic machine would fail to load a rivet and place it in position. This happens two or three times during a shift. He didnt me

23、ntion this in his earlier conversation with Jake because the problem had existed ever since he started working with this machine. The previous operator showed him how to clear the chute. All the rivet machines were like this according to Ben. Jake called the Machine Repair Department and asked that

24、someone look at the rivet track. A slight gap in the track was found and removed, and Ben continued to work. Delphi Confidential Two hours later, Jake got a call from Ben saying that the track was still jamming. As far as Jake could see, only rivets were in the bowl. Next, Jake looked into the rivet

25、 supplier containers. There was some foreign material in the blue container, but none in the red container. The label on the blue container showed that it was from Ajax Rivet, Inc., and the label on the red container indicated that it was from Franks Fasteners. Obviously, the foreign material was en

26、tering the rivet feeder bowl and jamming the track. Jake looked at the operator instructions posted on the job to see if there were any instructions about not using containers with foreign material in them, but there wasnt. Jake called Maintenance and requested that the bowl be cleaned. He also adde

27、d the cleaning operation to the preventive maintenance schedule on the equipment. He then called both Ajax Rivet, Inc. and Franks Fasteners. He asked about the cleaning procedures on the returnable containers. Franks did a full container purge and clean. Ajax just re-introduced the containers back i

28、nto their system. When Jake asked why Ajax did not clean their containers, he was told that Ajax was not aware that such a requirement was needed. It had never been part of the customer requirements document. 5 Why-Case Study Delphi Confidential 5 Why-Case Study Upon further investigation, Jake lear

29、ned that Franks Fasteners supplied other major automotive companies. Since these companies required that all returnable containers be cleaned, Franks instituted the purge as part of its practice for all customers. Ajax Rivet, however, depended primarily on Flex as its major customer. No such require

30、ment had ever been communicated to them. Jake called the Packaging Department and requested that a container maintenance requirement be drafted which would apply to all their suppliers. He also asked that a machine modification be developed to sense for the presence of rivets. Hopefully, this would

31、error proof the process. Delphi Confidential 5 Why-Case Study AN EXAMPLE - Part A A CUSTOMER REPORTED MISSING RIVETS IN A BRACKET ASSEMBLY. Automated rivet machine did not place rivet in position Why 1: Feeder had obstruction Why 2: Rivet supplier dunnage had foreign material Why 3: Supplier had no

32、maintenance program for dunnage Why 4: Supplier not aware of foreign material requirement Why 5: WHY DID THE PROBLEM OCCUR? Path to Investigate Specific Nonconformance Delphi Confidential 5 Why-Case Study AN EXAMPLE - Part B WHY WAS PROBLEM NOT DETECTED? No inspection required on Control Plan, excep

33、t set-up Why 1: Not identified on PFMEA as failure mode, except set-up Why 2: WE NOW KNOW WHY THE PROBLEM OCCURRED. BUT, IS THERE MORE? Path to Investigate Lack of Detection Delphi Confidential 5 Why-Case Study AN EXAMPLE - Part C WHY DID THE SYSTEM (QUALITY) ALLOW THIS TO HAPPEN? Potential failure

34、mode not identified on PFMEA Why 1: Incoming dunnage failure modes not addressed Why 2: Scope of PFMEA restricted to value add process Why 3: WE NOW KNOW WHY THE PROBLEM OCCURRED AND WHY IT WAS NOT DETECTED. BUT, IS THERE MORE? MISSING PIECE Path to Investigate system Delphi Confidential 5 Why-Case

35、Study Define Problem Use this path Missing rivets on brackets shipped to customer for the specific nonconformance being investigated WHY? Automated rivet machine did not place rivet in position WHY? Feedline had obstruction Use this path WHY? to investigate why the problem was not detected Foreign m

36、aterial found in dunnage obstructing feedline WHY? No requirement on control plan to sample brackets continuously; only at set-up Supplier had no maintenance program for dunnage WHY?WHY? Not identified on PFMEA as possible failure mode except at set-upB Use this path to investigate the systemic root

37、 cause Potential failure mode not identified on PFMEA WHY? Failure modes associated with incoming dunnage not addressed WHY? Scope of PFMEA limited to manufacturing process only C Corrective ActionsDate Root Causes A - Revise packing specification for dunnage to reflect no foreign material requireme

38、nt Supplier not aware of requirement about foreign material A B where and when it was observed, and where and when it might have been observed but was not; etc. -A useful tool for the first meeting of a problem solving team, because it puts the collective group knowledge on paper and shows what info

39、rmation is lacking and still needs to be collected. Why Is-Is Not Matrix? -Its the first step toward solving a problem systematically, because it lays out what is already known about the problem, in front of the entire problem solving team. -Its an efficient way in quick shooting the root cause. Del

40、phi Confidential Is-Is Not Matrix How to use Is-Is Not? -Clearly state the problem at hand, in terms of what the current state is versus what the desired state is. -Gather together a team consisting of people who would know something about the problem. -Fill out the Is-Is Not form in a group meeting

41、, at least initially. -Fill out the “Is Not” column as well as the “Is” column. -List any distinctions in the appropriate column. These distinctions can be useful clues. -List any changes that might have caused the problem to occur. Changes might be in supplier, raw material, machine, procedure, set

42、ting, etc. -If at any time, multiple problems seem to be appearing, it is OK to create separate Is-Is Not matrices for them. Tree diagram or cause-and-effect may also be used. -Develop action plan to collect any missing information, and to collect data that can back up any hearsay. Assign responsibl

43、e persons for that. -Analyze the completed matrix, try to find out clues. 1.If the problem is not yet solved, develop a problem solving plan based upon the Is-Is Not matrix. Try to incorporate other tools to find Red-X out. Delphi Confidential Is-Is Not Matrix ISIS NOTDistinctionsNeed more info WHAT

44、 object? What defect? WHERE geographically? Where on the object? WHEN in clock and calendar time? When since ? Any pattern? When in the history/life cycle? HOW MANY objects? How large is a defect? How many per object? What is the trend? Team Members: Delphi Confidential Is-Is Not Matrix-Example Sele

45、ct Contain Prevent Is Containment Necessary? No Yes Correct 1a. 問題描述問題描述 (站在客戶的角度看問題站在客戶的角度看問題)選擇選擇 當(dāng)前情況: 10月25日PSA中班3：50 質(zhì)量部質(zhì)檢員在包裝前的例行檢查時(shí)發(fā)現(xiàn)標(biāo)簽遺漏。 疑問點(diǎn)IS（是）（是）IS NOT （不是）（不是） DISTINCTIONS（區(qū)別）（區(qū)別） 問題是在哪里發(fā)生 的? PSA IPPSA ENG單板流水線 問題是在何時(shí)發(fā)生 的? 晚班早班1）分工操作不分工操作 2） 有領(lǐng)班無領(lǐng)班 發(fā)生的缺陷是什么?標(biāo)簽遺漏標(biāo)簽信息錯(cuò) 誤 遺漏是沒執(zhí)行某一操作，信息 錯(cuò)誤

46、是沒有正確執(zhí)行某一操作 首次發(fā)現(xiàn)問題是在 何時(shí)? 2003年第一 季度 2004年春節(jié) （開始開半 個(gè)班） 人員的熟練程度、工藝安排等 問題發(fā)生的頻率有 多大? 1年1次頻繁發(fā)生體系失效操作失效 Delphi Confidential Is-Is Not Matrix-Case Study Background Three pieces of Volvo Frond Door harness(P/N: 8697907) were rejected by VCG(Volvo Belgium Plant) respectively on Sep-23-2004, Oct-12-2004 and O

47、ct-15-2004. The reason for rejection is one connector can not be locked with the mating parts. When the final assembly operators at Volvo tried to mate these harness with mating parts, a clear “click” sound could not be heard, and the mating parts could be easily pulled away from each others, which

48、indicated a non-locked connection. The not-locked situation existed after disassembling and reassembling multiple times. The defective harness was originally built at Delphi GZ, then shipped to Delphi Europe, then transported to Volvos plants. Delphi GZ is building both Volvo Front Door harness and

49、Rear Door harness. The not locked connector mentioned above is used in both front door harness and rear door harness, one each. Sorting has been conducted at Delphi Europe for all harness in stock. Totally 50 pieces defective harness were found. Directions -Fill out an initial Is-Is Not form based u

50、pon the above information -Any other information needed? -Do not try to solve technical problems in the case study. Delphi Confidential Problem Solving Tool box - Problem Solving Tree Delphi Confidential Problem Solving Tree Problem Description Possible Reason possible reason possible reason possibl

51、e reason possible reason possible reason possible reason possible reason possible reason possible reason possible reason possible reason possible reason possible reason possible reason possible reason possible reason possible reason possible reason possible reason possible reason possible reason pos

52、sible reason possible reason possible reason Root Cause Possible Reason Possible Reason A Problem Solving Tree looks like this Delphi Confidential Problem Solving Tree Whats Problem Solving Tree? -A tree diagram used to break up a vaguely stated problem into specific source of causes and finally aim

53、s at root cause by crossing-off those that do not matter. It is similar to Shainins Problem Definition Treesm, Project Definition Treesm, and Solution Treesm. -A tree diagram that is mapped down by continuously asking “why” and thinking in the way of “Is-Is Not”. -A useful tool for the first meeting

54、 and routine meetings of a problem solving team, because it puts the collective group knowledge on paper and shows what information is lacking and still needs to be collected. -A systematic way of problem solving by dividing sources of possible causes into logic families, which is powerful and favor

55、ed by our customers Delphi Confidential Problem Solving Tree How to map a Problem Solving Tree? -Start with a summarized problem description, which actually is a project definition. -Map the first level of families of possible sources of causes. -Break up first level of families into second level of

56、 families. -Continue above step until the root cause is found. -Make splits. (Cross-off those families that are not causes of the specific problem, by using any appropriate tools, including Shainin techniques, DPS tools, 6-Sigma, etc.) -Finally point at the root cause. -Confirm the identity of root cause. 1.Wrap up the investigation. Note: Rome was not built in one day. Dont try to find out the root cause at the first