解SNa的 ssn meangnpneeara民n, anaD NDT alapat SN Lrdl trrh sirum ngale BSN17Su兵 nade lw pelerin Haht enton- Ci)ss A wAlUT t加1an,和 are sDf malan!n ad retrieal sN2DaG引da an AAA bel uT r ard Not MoNto poseurs NDT easton 向14s2 上筐 ulen rhino: NDT e I ENI Deed Catyhad nach s rdson and SN 2 Dact CaaN Mes 二需需蕊 Fig. 3 Example of a failure mode assessment chart (for fault tree of forgings defects in Decision-making protocols can be as varied as the different kinds of decisions. Uncertainty and risk are part of most decisions, and so some assessment of probability is needed. As part of a failure investigation, first assess the probability of each potential root cause. How realistic is it? Each root cause is given a probability, such as likely, possible, or unlikely. Second, assign a priority to each potential root cause. This priority establishes the order in which the failure investigation will evaluate each root cause. Factors that effect this decision are hardware availability, amount of available hardware for test, cost, affect of the planned test on other tests whether one test can be used to disprove or prove multiple root causes, and so on. Third, document the rationale used to assign a root cause its particular probability and priority. Once the failure investigation is started, no one will remember why a certain test was performed first or last if this rationale is not documented. Any action items directed toward proving or disproving the root cause should be listed in an abbreviated manner There are many reasons an FMA chart is important. The FMA chart is a permanent record of the assessed probability and assigned priority of each root cause, combined with the rationale for the particular assessments The FMA chart also lists the action items to prove or disprove each root cause in an abbreviated form. Once again, this chart is excellent for briefings. The failure investigation is now documented on two easy charts, the fault tree and the fma Similar to the fault tree, the Fma chart is a living list that can be changed at any time. Once the failure investigation is in progress, information is generated that proves or disproves each root cause. The information may also lead to a new probability or priority rating. a quick glance at this chart informs the investigator what happens if the probability or priority of one root cause is changed. The priority listing in the FMA provides th order in which each root cause is proved or disproved. This leads directly to the next step Step 5: Converge on the Most Likely Root Cause(s). Now it is time to converge on the most likely root causes(s). This would be step 2 in the process listed in the previous section, " The Four-Step Problem-Solving Process. A good tool for this is the Technical Plan for Resolution(TPR)chart(Fig 4), which can be created with a spreadsheet program similar to the FMa chart. In fact, a good practice is to tie the FMA and TPR charts Thefileisdownloadedfromwww.bzfxw.com
Fig. 3 Example of a failure mode assessment chart (for fault tree of forgings defects in Fig. 2) Decision-making protocols can be as varied as the different kinds of decisions. Uncertainty and risk are part of most decisions, and so some assessment of probability is needed. As part of a failure investigation, first assess the probability of each potential root cause. How realistic is it? Each root cause is given a probability, such as likely, possible, or unlikely. Second, assign a priority to each potential root cause. This priority establishes the order in which the failure investigation will evaluate each root cause. Factors that effect this decision are hardware availability, amount of available hardware for test, cost, affect of the planned test on other tests, whether one test can be used to disprove or prove multiple root causes, and so on. Third, document the rationale used to assign a root cause its particular probability and priority. Once the failure investigation is started, no one will remember why a certain test was performed first or last if this rationale is not documented. Any action items directed toward proving or disproving the root cause should be listed in an abbreviated manner. There are many reasons an FMA chart is important. The FMA chart is a permanent record of the assessed probability and assigned priority of each root cause, combined with the rationale for the particular assessments. The FMA chart also lists the action items to prove or disprove each root cause in an abbreviated form. Once again, this chart is excellent for briefings. The failure investigation is now documented on two easy charts, the fault tree and the FMA. Similar to the fault tree, the FMA chart is a living list that can be changed at any time. Once the failure investigation is in progress, information is generated that proves or disproves each root cause. The information may also lead to a new probability or priority rating. A quick glance at this chart informs the investigator what happens if the probability or priority of one root cause is changed. The priority listing in the FMA provides the order in which each root cause is proved or disproved. This leads directly to the next step. Step 5: Converge on the Most Likely Root Cause(s). Now it is time to converge on the most likely root causes(s). This would be step 2 in the process listed in the previous section, “The Four-Step Problem-Solving Process.” A good tool for this is the Technical Plan for Resolution (TPR) chart (Fig. 4), which can be created with a spreadsheet program similar to the FMA chart. In fact, a good practice is to tie the FMA and TPR charts The file is downloaded from www.bzfxw.com
together on two sheets within one spreadsheet file, so that any changes made to the Fma are immediately reflected in the TPR mbl Faclty 3 Enpinewrrg 升 ew Supper B caving Pro的9 neral and he hyper contant Faur CaO sNared panora and ha we la Work ewen rype d N 下■■ Fig. 4 Example of a technical plan for resolution Previous steps (3 and 4)have provided the root causes, assessed the probability, and assigned the priority each root cause. Now comes the time to create the technical plan to prove or disprove each root cause. The TPr
together on two sheets within one spreadsheet file, so that any changes made to the FMA are immediately reflected in the TPR. Fig. 4 Example of a technical plan for resolution Previous steps (3 and 4) have provided the root causes, assessed the probability, and assigned the priority of each root cause. Now comes the time to create the technical plan to prove or disprove each root cause. The TPR
ensures that the testing or analysis is performed in a manner that should achieve maximum effect and efficiency. It is a detailed road map in which one can independently list the work needed to prove or disprove each root cause. Once the details are down on paper, a specific order of testing or analysis or a synergy of investigator may discover he runs out of material by the time he gets to a certain test or analysis At this stage, a good tool is to ask the investigative team three questions about the failure What do i know? What do i think What can I prove? What do I know? These are the hard facts and existing data. List the facts and then determine to which root ause(s)they are applicable. When the list of facts is completed, ask"What do i think? These are the theories, eas. an nd possible cause-and-effect relationships. These theories and ideas are based on history, experience logic, and the hard facts noted previously. This is the time for personnel to ask"What if...? Finally, ask"What can I prove?" For each of the theories, ideas, or root causes, list the physical evidence one would expect to see if it occurred, as well as the tests, analyses, and so on required to prove or disprove each one. It is sometimes easier to disprove a theory for the root cause than to prove it. This may involve testing or analysis of the failed hardware, testing or analysis of similar hardware, exemplar testing to demonstrate if a failure mode is possible, operational data analysis, stress analysis, literature searches, a phone call to other divisions or companies that might have seen a similar problem, and so on It is equally important to list who has the action to get a test or analysis performed, the completion date, and the results on the TPR. It cannot be stressed enough that the person assigned must agree both to performing the test or analysis and to the completion date. Just assigning a task does not always ensure that it gets done, but if someone agrees to the task and the deadline, then the work normally gets accomplished. Documenting the results of the test plus any comments on the TPR completes the documentation Similar to the Fault Tree and Fma chart, there are many reasons a tPR chart is important. First and foremost, the TPR chart is a permanent record of each test or analysis that is to be performed, the person assigned, the expected completion date, and the results, with any comments. Once again, this chart is excellent for briefings In addition, if any changes occur on the Fault Tree and FMa chart, the tpr chart will indicate whether there is a need for more testing or if the necessary testing is already planned Step 6: Objectively and Clearly Identify All Possible Corrective Actions. The failure investigation is not completed yet. In fact, it is just half done. After weeks or months of hard work, the failure investigation determines the root cause(s). Unfortunately, many failure investigations end here, but now the investigator needs to determine the corrective action(s)to prevent the failure from occurring again In this step, all possible corrective actions are identified This is step 3 from the Four-Step Problem-Solving Process, "What are the potential solutions? Step 6 is similar in concept to Step 3, Objectively and clearly identify all possible root causes. " The difference is that the failure investigation team is now centered on identifying all possible corrective actions to solve the root cause identified in the failure analysis. Repeat the instructions in step 3, and create a Corrective Action Tree(Fig. 5) Thefileisdownloadedfromwww.bzfxw.com
ensures that the testing or analysis is performed in a manner that should achieve maximum effect and efficiency. It is a detailed road map in which one can independently list the work needed to prove or disprove each root cause. Once the details are down on paper, a specific order of testing or analysis or a synergy of testing or analysis becomes evident. The investigator may change the priority of some testing or analysis, because it shows up under multiple root causes and is a good test to perform early. Alternatively, the investigator may discover he runs out of material by the time he gets to a certain test or analysis. At this stage, a good tool is to ask the investigative team three questions about the failure: · What do I know? · What do I think? · What can I prove? “What do I know?” These are the hard facts and existing data. List the facts and then determine to which root cause(s) they are applicable. When the list of facts is completed, ask “What do I think?” These are the theories, ideas, and possible cause-and-effect relationships. These theories and ideas are based on history, experience, logic, and the hard facts noted previously. This is the time for personnel to ask “What if…?” Finally, ask “What can I prove?” For each of the theories, ideas, or root causes, list the physical evidence one would expect to see if it occurred, as well as the tests, analyses, and so on required to prove or disprove each one. It is sometimes easier to disprove a theory for the root cause than to prove it. This may involve testing or analysis of the failed hardware, testing or analysis of similar hardware, exemplar testing to demonstrate if a failure mode is possible, operational data analysis, stress analysis, literature searches, a phone call to other divisions or companies that might have seen a similar problem, and so on. It is equally important to list who has the action to get a test or analysis performed, the completion date, and the results on the TPR. It cannot be stressed enough that the person assigned must agree both to performing the test or analysis and to the completion date. Just assigning a task does not always ensure that it gets done, but if someone agrees to the task and the deadline, then the work normally gets accomplished. Documenting the results of the test plus any comments on the TPR completes the documentation. Similar to the Fault Tree and FMA chart, there are many reasons a TPR chart is important. First and foremost, the TPR chart is a permanent record of each test or analysis that is to be performed, the person assigned, the expected completion date, and the results, with any comments. Once again, this chart is excellent for briefings. In addition, if any changes occur on the Fault Tree and FMA chart, the TPR chart will indicate whether there is a need for more testing or if the necessary testing is already planned. Step 6: Objectively and Clearly Identify All Possible Corrective Actions. The failure investigation is not completed yet. In fact, it is just half done. After weeks or months of hard work, the failure investigation determines the root cause(s). Unfortunately, many failure investigations end here, but now the investigator needs to determine the corrective action(s) to prevent the failure from occurring again. In this step, all possible corrective actions are identified. This is step 3 from the “Four-Step Problem-Solving Process,” “What are the potential solutions?” Step 6 is similar in concept to Step 3, “Objectively and clearly identify all possible root causes.” The difference is that the failure investigation team is now centered on identifying all possible corrective actions to solve the root cause identified in the failure analysis. Repeat the instructions in step 3, and create a Corrective Action Tree (Fig. 5). The file is downloaded from www.bzfxw.com
What is the Corrective Action for the Forgings with Penetrant Defects? Casting Process Forging Process Heat Treatment Improvements Improvements Process Improvements Nonmetallic Shrinked H2 i Content Forging Reduction Reduction Reduction Reduction 率 Develop Ing Slice Exam all Ingots and H2 content Reduction Penetrant or Inspection of Inspection Casting (201 and 3A1) Dovice 1A1) (2A1) Improve Forging Strain Rate (2C2 and 3A2) (2A2 41) Add Fluxing an Improve Forging Strain Direction 1A2A Fig 5 Example of corrective action tree for forgings with dye-penetrant defects. LIMCA liquid metal cleanness analyzer device Step 7: Objectively Evaluate Each Corrective Action. The next step of the failure investigation is to objectively evaluate the likelihood of each corrective action listed in the corrective action tree. This step is a repeat of step 4, "Objectively evaluate likelihood of each root cause. The difference is that the failure investigation team is assessing the probability and assigning priority to each corrective action, combined with the rationale. Repeat he instructions in step 4, and create a Corrective Action Assessment(CAA)chart Step 8: Select Optimal Corrective Action(s). Now it is time to select the optimal corrective action(s). This is step 4 from the Four-Step Problem-Solving Process, What is the best solution? This step is a repeat of step 5, Converge on the most likely root cause(s). At this step, the failure analysis team creates the technical plan to determine the best corrective action(s). Repeat the instructions in step 5, and create a Technical Plan for Evaluation(tPe)chart Step 9: Evaluate Effectiveness of Selected Corrective Action(s). The last step is to recommend an evaluation of the corrective action(s) selected. After some period of time, an evaluation of the corrective action should be performed. Whether the evaluation should be performed once or on a periodic basis is at the discretion of the failure investigation team. It is also important that the evaluation be a proper test of the corrective action Summary The goal of any failure investigation is to discover the root cause(s). The first five steps of the nine-step process described here point the failure investigator toward that goal. Once the root cause(s)is determined, the second part of a failure investigation is to determine the best corrective action(s), as noted in the last four steps The three basic tools identified in this article that are helpful in any failure investigation are a Fault Tree, an FMA chart, and a TPR chart. These three tools provide a documented, interchangeable, and concise set of
Fig. 5 Example of corrective action tree for forgings with dye-penetrant defects. LIMCA, liquid metal cleanness analyzer device Step 7: Objectively Evaluate Each Corrective Action. The next step of the failure investigation is to objectively evaluate the likelihood of each corrective action listed in the corrective action tree. This step is a repeat of step 4, “Objectively evaluate likelihood of each root cause.” The difference is that the failure investigation team is assessing the probability and assigning priority to each corrective action, combined with the rationale. Repeat the instructions in step 4, and create a Corrective Action Assessment (CAA) chart. Step 8: Select Optimal Corrective Action(s). Now it is time to select the optimal corrective action(s). This is step 4 from the “Four-Step Problem-Solving Process,” “What is the best solution?” This step is a repeat of step 5, “Converge on the most likely root cause(s).” At this step, the failure analysis team creates the technical plan to determine the best corrective action(s). Repeat the instructions in step 5, and create a Technical Plan for Evaluation (TPE) chart. Step 9: Evaluate Effectiveness of Selected Corrective Action(s). The last step is to recommend an evaluation of the corrective action(s) selected. After some period of time, an evaluation of the corrective action should be performed. Whether the evaluation should be performed once or on a periodic basis is at the discretion of the failure investigation team. It is also important that the evaluation be a proper test of the corrective action. Summary The goal of any failure investigation is to discover the root cause(s). The first five steps of the nine-step process described here point the failure investigator toward that goal. Once the root cause(s) is determined, the second part of a failure investigation is to determine the best corrective action(s), as noted in the last four steps discussed. The three basic tools identified in this article that are helpful in any failure investigation are a Fault Tree, an FMA chart, and a TPR chart. These three tools provide a documented, interchangeable, and concise set of
information to ensure that all possible root causes are evaluated in the most efficient manner and order. These three tools are then repeated for the corrective actions in the form of the Corrective Action Tree, the caa chart and the TPe chart Organization of a Failure Investigation Daniel P Dennies, Ph. D, The Boeing Company Failure lnvestigation pitfalls However, there are pitfalls along the way. Common problems and mistakes are The magic bullet or silver bullet theory Not understanding the problem Not considering all possible failure causes Jumping to conclusions Not identifying all the root cause(s) Remove and replace"company mentality Not asking for help Failing to follow through Not understanding how the failed system is supposed to operate Tearing the system apart without a developed plan Some companies believe there are substitutes or shortcuts for a root-cause failure investigation that can save time and money In the end these substitutes do neither These substitutes include Give me your best guess Give me a five minute failure analysis Rework and repair based on no failure investigation Return the part to the supplier and let them figure it out Scrapping the hardware Swapping out for another part Ignoring the problem and hoping it is a unique occurrence "Band-aid" fixes that do not address the root cause In the end, these substitutes are generally neither cost-effective nor schedule friendly. Furthermore, if a failure involves personal injury, property damage, or other economic loss, the failure to find the root cause as soon as possible may expose a company to further liability Organization of a Failure Investigation Daniel P. Dennies, Ph D. The Boeing Company Other tools Every failure investigation is unique. Therefore, it pays to have many tools in your arsenal. This article presents one set of tools for the organization of a failure investigation There are other tools as well but the author has developed this approach and finds it to be a cost-effective technique that works extremely well in most situations Thefileisdownloadedfromwww.bzfxw.com
information to ensure that all possible root causes are evaluated in the most efficient manner and order. These three tools are then repeated for the corrective actions in the form of the Corrective Action Tree, the CAA chart, and the TPE chart. Organization of a Failure Investigation Daniel P. Dennies, Ph.D., The Boeing Company Failure Investigation Pitfalls However, there are pitfalls along the way. Common problems and mistakes are: · The magic bullet or silver bullet theory · Not understanding the problem · Not considering all possible failure causes · Jumping to conclusions · Not identifying all the root cause(s) · “Remove and replace” company mentality · Not asking for help · Failing to follow through · Not understanding how the failed system is supposed to operate · Tearing the system apart without a developed plan Some companies believe there are substitutes or shortcuts for a root-cause failure investigation that can save time and money. In the end, these substitutes do neither. These substitutes include: · “Give me your best guess” · “Give me a five minute failure analysis” · Rework and repair based on no failure investigation · Return the part to the supplier and let them figure it out · Scrapping the hardware · Swapping out for another part · Ignoring the problem and hoping it is a unique occurrence · “Band-aid” fixes that do not address the root cause In the end, these substitutes are generally neither cost-effective nor schedule friendly. Furthermore, if a failure involves personal injury, property damage, or other economic loss, the failure to find the root cause as soon as possible may expose a company to further liability. Organization of a Failure Investigation Daniel P. Dennies, Ph.D., The Boeing Company Other Tools Every failure investigation is unique. Therefore, it pays to have many tools in your arsenal. This article presents one set of tools for the organization of a failure investigation. There are other tools as well, but the author has developed this approach and finds it to be a cost-effective technique that works extremely well in most situations. The file is downloaded from www.bzfxw.com