Understanding how the typical distribution of failure for a given product must be factored with time is also important when looking at failure patterns(Fig. 3). Early life failures are often associated with fabrication issues, quality-control issues. or initial"shakedown"stresses. while age-related failure rates would increase with time. This is discussed in more detail in the article "Reliability-Centered Maintenance"in this volume failure Wearout failure period Intrinsic failure period Time Fig3 Typical time distribution of failures("bathtub curve Once the concept of a managed life is prudently adopted over a simple failure prevention concept, design and fabrication costs can be reduced and maintenance and other life-prolonging activities can be optimized Diligence in Use of Terminology. Communicating technical information is of paramount importance in all engineering areas, including failure analysis. The choice of technical descriptors, nomenclature, and even what might be considered technical jargon is critical to conveying technical ideas to other engineers, managers, plant personnel, shop personnel maintenance personnel, attorneys, a jury, and so forth. It is instructive in this introductory article to emphasize that a descriptor can mean something very specific to a technical person and mean something very different to a business manager or an attorne For example, the term"flaw is synonymous with defect"in general usage. However, to a fracture mechanics specialist, flaw is a discontinuity such as a crack. Under some circumstances, when the crack is smaller than the critical size (i.e subcritical), the crack is benign and therefore may not be considered a defect. To the quality-control engineer, flaws are characteristics that are managed continuously on the production line, as every engineered product has flaws, or deviations from perfection"(Ref 14). On the manufacturing floor, these flaws are measured, compared with the preestablished limits of acceptability, and dispositioned as acceptable or rejectable. A rejectable characteristic is defined as a defect(Ref 14). To the Six Sigma practioner, a defect is considered anything that inhibits a process or, in a broad sense, any condition that fails to meet a customer expectation(Ref 9). To the attorney, a defect refers to many different types of deficiencies, including improper design, inadequate instructions for use, insufficient warnings, and even inappropriate advertising or marketing(Ref 15) Similar nuances may occur in the basic definitions and interpretations of technical terms used in materials failure analysis Terms such as ductile and brittle, crack and fracture, and stable and unstable crack growth are pervasive in failure analysis. Even these seemingly basic terms are subject to misuse and misinterpretations, as suggested in Ref 16--for example"brittle cleavage, which is a pleonasm that does not explain anything. Another example noted in Ref 16 is the term " overload fracture, which may be misinterpreted by nonanalysts as a failure caused by a load higher anticipated by the materials or mechanical engineers. This limited interpretation of overload failure is incomplete described in the article"Overload Failures"in this volume Judgmental terminology should be used with prudence when communicating analytical protocols, procedures, findings, and conclusions. Communications during the preliminary stages of an investigation should be factual rather than judgmental. It is important to recognize that some of the terminology used in a failure analysis can be judgmental, and onsideration must be given to the implications associated with the use of such terminology. For example, when examining both a failed and an unfailed component returned from service, references to the unfailed sample as"good and the failed sample as"bad"should be avoided. This is because the investigation may reveal both samples to contain the same defect, and therefore both could be considered"bad. Similarly, neither may be bad" if the analysis actually indicates the failed component met all requirements but was subjected to abuse in service. On completion of the failure analysis, judgmental terminology is often appropriate to use if the evidence supports it, such as in the example of asting defect that has been confirmed in the example bolt failure analysis While discussions of the semantics of terminology may seem pedantic, communicating the intended information gleaned from a failure analysis relies heavily on precision in the use of language References cited in this section 4. W.E. Deming, Out of the Crisis, MIT Center for Advanced Engineering Study, 1986
Understanding how the typical distribution of failure for a given product must be factored with time is also important when looking at failure patterns (Fig. 3). Early life failures are often associated with fabrication issues, quality-control issues, or initial “shakedown” stresses, while age-related failure rates would increase with time. This is discussed in more detail in the article “Reliability-Centered Maintenance” in this Volume. Fig. 3 Typical time distribution of failures (“bathtub curve”) Once the concept of a managed life is prudently adopted over a simple failure prevention concept, design and fabrication costs can be reduced and maintenance and other life-prolonging activities can be optimized. Diligence in Use of Terminology. Communicating technical information is of paramount importance in all engineering areas, including failure analysis. The choice of technical descriptors, nomenclature, and even what might be considered technical jargon is critical to conveying technical ideas to other engineers, managers, plant personnel, shop personnel, maintenance personnel, attorneys, a jury, and so forth. It is instructive in this introductory article to emphasize that a descriptor can mean something very specific to a technical person and mean something very different to a business manager or an attorney. For example, the term “flaw” is synonymous with “defect” in general usage. However, to a fracture mechanics specialist, a flaw is a discontinuity such as a crack. Under some circumstances, when the crack is smaller than the critical size (i.e., subcritical), the crack is benign and therefore may not be considered a defect. To the quality-control engineer, flaws are characteristics that are managed continuously on the production line, as every engineered product has flaws, or “deviations from perfection” (Ref 14). On the manufacturing floor, these flaws are measured, compared with the preestablished limits of acceptability, and dispositioned as acceptable or rejectable. A rejectable characteristic is defined as a defect (Ref 14). To the Six Sigma practioner, a defect is considered anything that inhibits a process or, in a broad sense, any condition that fails to meet a customer expectation (Ref 9). To the attorney, a defect refers to many different types of deficiencies, including improper design, inadequate instructions for use, insufficient warnings, and even inappropriate advertising or marketing (Ref 15). Similar nuances may occur in the basic definitions and interpretations of technical terms used in materials failure analysis. Terms such as ductile and brittle, crack and fracture, and stable and unstable crack growth are pervasive in failure analysis. Even these seemingly basic terms are subject to misuse and misinterpretations, as suggested in Ref 16—for example “brittle cleavage,” which is a pleonasm that does not explain anything. Another example noted in Ref 16 is the term “overload fracture,” which may be misinterpreted by nonanalysts as a failure caused by a load higher than anticipated by the materials or mechanical engineers. This limited interpretation of overload failure is incomplete, as described in the article “Overload Failures” in this Volume. Judgmental terminology should be used with prudence when communicating analytical protocols, procedures, findings, and conclusions. Communications during the preliminary stages of an investigation should be factual rather than judgmental. It is important to recognize that some of the terminology used in a failure analysis can be judgmental, and consideration must be given to the implications associated with the use of such terminology. For example, when examining both a failed and an unfailed component returned from service, references to the unfailed sample as “good” and the failed sample as “bad” should be avoided. This is because the investigation may reveal both samples to contain the same defect, and therefore both could be considered “bad.” Similarly, neither may be “bad” if the analysis actually indicates the failed component met all requirements but was subjected to abuse in service. On completion of the failure analysis, judgmental terminology is often appropriate to use if the evidence supports it, such as in the example of a casting defect that has been confirmed in the example bolt failure analysis. While discussions of the semantics of terminology may seem pedantic, communicating the intended information gleaned from a failure analysis relies heavily on precision in the use of language. References cited in this section 4. W.E. Deming, Out of the Crisis, MIT Center for Advanced Engineering Study, 1986
5. J M. Juran and F, M. Gryna, Ed. Juran's Quality Control Handbook, 4th ed, McGraw-Hill, 1988 6. P.F. Wilson, L D. Dell, and G F. Anderson, Root Cause Analysis: A Tool for Total Quality Management, ASQ Quality Press, 1993, p 7 7. F.w. Breyfogle Ill, Implementing Six Sigma: Smarter Solutions Using Statistical Methods, John Wiley Sons 1999, p xxvii 8. P.S. Pande, R. P. Neuman, and R.R. Cavanaugh, The Six Sigma Way, McGraw-Hill, 2000, P xi 9. M. Harry and R. Schroeder, Six Sigma: The Breakthrough Management Strategy Revolutionizing the World's Top Corporations, Doubleday Co, Inc, 1999 10. G.F. Smith, Quality Problem Solving, AsQ Quality Press, 1998, p7 11. B. Anderson and T. Fagerhaug, Root Cause Analysis: Simplified Tools and Techniques, ASQ Quality Press, 2000 p7,125 Reporting Workplace Errors, Max Ammerman/Quality Resources, 1998 oach to Identifying, Correcting, and 12. M. Ammerman, The Root Cause Analysis Handbook: A Simplified Appr 13. Engineering Aspects of Failure and Failure Analysis, Failure Analysis and Prevention, Vol 10, &th ed. Metals Handbook, American Society for Metals, 1975, p 1-9 14. R.K. McLeod, T Heaslip, and M. Vermij, Defect or Flaw-Legal Implications, Failure Analysis: Techniques and Applications, Conf. Proc. International Conference and Exhibits on Failure Analysis, 8-11 July 1991(Montreal Quebec, Canada), ASM International, 1992, p 253-261 15. J.J. Asperger, Legal Definition of a Product Failure: What the Law Requires of the Designer and the Manufacturer, Proc. Failure Prevention through Education: Getting to the Root Cause, 23-25 May 2000 ( Cleveland, OH), ASM International, 2000, p 25-29 6. D. Broek, Fracture Mechanics as an Important Tool in Failure Analysis, Failure Analysis: Techniques Applications, Conf. Proc. International Conference and Exhibits on Failure Analysis, 8-11 July 1991(Mont Quebec, Canada), ASM International, 1992, p 33-44 Introduction to Failure Analysis and Prevention James ]. Scutti, Massachusetts Materials Research, Inc. William ]. McBrine, ALTRAN Corporation Root-Cause analysi Failure analysis is considered to be the examination of the characteristics and causes of equipment or component failure In most cases this involves the consideration of physical evidence and the use of engineering and scientific principles and analytical tools. Often, the reason why one performs a failure analysis is to characterize the causes of failure with the erall objective to avoid repeat of similar failures. However, analysis of the physical evidence alone may not be adequate to reach this goal. The scope of a failure analysis can, but does not necessarily, lead to a correctable root cause of failure. Many times, a failure analysis incorrectly ends at the identification of the failure mechanism and perhaps cau influences. The principles of root-cause analysis(RCA)may be applied to ensure that the root cause is understood and appropriate corrective actions may be identified. An RCa exercise may simply be a momentary mental exercise or an extensive logistical charting analysis Many volumes have been written on the process and methods of RCA. The concept of rca does not apply to failures alone, but is applied in response to an undesirable event or condition(Fig. 4). Root-cause analysis is intended to identify the fundamental cause(s) that if corrected will prevent recurrence Thefileisdownloadedfromwww.bzfxw.com
5. J.M. Juran and F.M. Gryna, Ed., Juran's Quality Control Handbook, 4th ed., McGraw-Hill, 1988 6. P.F. Wilson, L.D. Dell, and G.F. Anderson, Root Cause Analysis: A Tool for Total Quality Management, ASQ Quality Press, 1993, p 7 7. F.W. Breyfogle III, Implementing Six Sigma: Smarter Solutions Using Statistical Methods, John Wiley & Sons, 1999, p xxvii 8. P.S. Pande, R.P. Neuman, and R.R. Cavanaugh, The Six Sigma Way, McGraw-Hill, 2000, p xi 9. M. Harry and R. Schroeder, Six Sigma: The Breakthrough Management Strategy Revolutionizing the World's Top Corporations, Doubleday & Co., Inc., 1999 10. G.F. Smith, Quality Problem Solving, ASQ Quality Press, 1998, p 7 11. B. Anderson and T. Fagerhaug, Root Cause Analysis: Simplified Tools and Techniques, ASQ Quality Press, 2000, p 7, 125 12. M. Ammerman, The Root Cause Analysis Handbook: A Simplified Approach to Identifying, Correcting, and Reporting Workplace Errors, Max Ammerman/Quality Resources, 1998 13. Engineering Aspects of Failure and Failure Analysis, Failure Analysis and Prevention, Vol 10, 8th ed., Metals Handbook, American Society for Metals, 1975, p 1–9 14. R.K. McLeod, T. Heaslip, and M. Vermij, Defect or Flaw—Legal Implications, Failure Analysis: Techniques and Applications, Conf. Proc. International Conference and Exhibits on Failure Analysis, 8–11 July 1991 (Montreal, Quebec, Canada), ASM International, 1992, p 253–261 15. J.J. Asperger, Legal Definition of a Product Failure: What the Law Requires of the Designer and the Manufacturer, Proc. Failure Prevention through Education: Getting to the Root Cause, 23–25 May 2000 (Cleveland, OH), ASM International, 2000, p 25–29 16. D. Broek, Fracture Mechanics as an Important Tool in Failure Analysis, Failure Analysis: Techniques and Applications, Conf. Proc. International Conference and Exhibits on Failure Analysis, 8–11 July 1991 (Montreal, Quebec, Canada), ASM International, 1992, p 33–44 Introduction to Failure Analysis and Prevention James J. Scutti, Massachusetts Materials Research, Inc.; William J. McBrine, ALTRAN Corporation Root-Cause Analysis Failure analysis is considered to be the examination of the characteristics and causes of equipment or component failure. In most cases this involves the consideration of physical evidence and the use of engineering and scientific principles and analytical tools. Often, the reason why one performs a failure analysis is to characterize the causes of failure with the overall objective to avoid repeat of similar failures. However, analysis of the physical evidence alone may not be adequate to reach this goal. The scope of a failure analysis can, but does not necessarily, lead to a correctable root cause of failure. Many times, a failure analysis incorrectly ends at the identification of the failure mechanism and perhaps causal influences. The principles of root-cause analysis (RCA) may be applied to ensure that the root cause is understood and appropriate corrective actions may be identified. An RCA exercise may simply be a momentary mental exercise or an extensive logistical charting analysis. Many volumes have been written on the process and methods of RCA. The concept of RCA does not apply to failures alone, but is applied in response to an undesirable event or condition (Fig. 4). Root-cause analysis is intended to identify the fundamental cause(s) that if corrected will prevent recurrence. The file is downloaded from www.bzfxw.com
Systems and indicators Evident cause(s) Hidden root Fig 4 Root-cause analogy Levels. The three levels of root-cause analysis are physical roots, human roots, and latent roots(Ref 17, 18, 19, 20, 21) Physical roots, or the roots of equipment problems, are where many failure analyses stop These roots may be what comes out of a laboratory investigation or engineering analysis and are often component-level or materials-level findings Human roots (i.e, people issues) involve human factors that caused the failure, an example being an error in human judgment. Latent roots lead us to the causes of the human error and include roots that are organizational or procedural in nature. as well as environmental or other roots that are outside the realm of control These levels or root cause are best defined by the two examples Table 2 Examples of root causes of failure of pressure vessel and bolt root type Pressure vessel fail Bolt failure Physical roots Corrosion damage, wall thinning Fatigue crack; equipment vibration. lack of vibration;isolation Human roots Inadequate inspection performed Improper equipment installed Latent roots nadequate inspector training Inadequate specification verification process How deeply one goes into the root causes depends on the objectives of the RCA. These objectives are typically based on the complexity of the situation and the risk associated with additional failures. In most cases, one desires to identify root causes that are reasonably correctable. An example of the variety of possible root causes of an electric motor driven compressor assembly is provided in Table 3(Ref 22) Table 3 Possible causes of electric motor driven pump or compressor failures System design Component Shipping and Installation Operations and Distress damage or and specification manufacturer's storage responsibility maintenance failed components responsibility responsibility esponsibility responsibilit Application Material of Preparation Foundations Shock Distress damages construction for shipment Undercapacity Settling Thermal Vibration Overcapacity Improper materia//clsystem not Flaw or defect Improper or Mechanical Short/open circuit sufficient Inadequ tartup Failed components
Fig. 4 Root-cause analogy Levels. The three levels of root-cause analysis are physical roots, human roots, and latent roots (Ref 17, 18, 19, 20, 21). Physical roots, or the roots of equipment problems, are where many failure analyses stop. These roots may be what comes out of a laboratory investigation or engineering analysis and are often component-level or materials-level findings. Human roots (i.e., people issues) involve human factors that caused the failure, an example being an error in human judgment. Latent roots lead us to the causes of the human error and include roots that are organizational or procedural in nature, as well as environmental or other roots that are outside the realm of control. These levels or root cause are best defined by the two examples in Table 2. Table 2 Examples of root causes of failure of pressure vessel and bolt Root type Pressure vessel failure Bolt failure Physical roots Corrosion damage, wall thinning Fatigue crack; equipment vibration; lack of vibration; isolation Human roots Inadequate inspection performed Improper equipment installed Latent roots Inadequate inspector training Inadequate specification verification process How deeply one goes into the root causes depends on the objectives of the RCA. These objectives are typically based on the complexity of the situation and the risk associated with additional failures. In most cases, one desires to identify root causes that are reasonably correctable. An example of the variety of possible root causes of an electric motor driven compressor assembly is provided in Table 3 (Ref 22). Table 3 Possible causes of electric motor driven pump or compressor failures System design and specification responsibility Component manufacturer's responsibility Shipping and storage responsibility Installation responsibility Operations and maintenance responsibility Distress damage or failed components Application Undercapacity Overcapacity Incorrect physical Material of construction Flaw or defect Improper material Preparation for shipment Oil system not clean Inadequate Foundations Settling Improper or insufficient grouting Shock Thermal Mechanical Improper startup Distress damages Vibration Short/open circuit Failed components
System design Component Operations and Distress damage or and specification manufacturer's storage responsibility maintenance failed components responsibility responsibility responsibility responsibility condition assumed Improper drainage (temperature Cracking or Operating Sleeve bearing pressure, etc.) Design Incorrect physical applie Piping property assumedImproper Process surging Coupling (molecular weight, specification Wrong coating Misalignment Control error Shaft Wrong selection Inadequate Specifications Equipment not cleaning Controls Pinion/ball/turning Design error deactivated/not Inadequate installed lubrication system Inadequate or Protection support Casir wrong lubrication Operating error Insufficient control Insufficient Assembly Rotor instrumentation Inadequate liquid protection Auxiliaries drain Misalignment Impeller Improper coupling Corrosion by Utility failure Critical speed salt Assembly Shroud Improper bearing amage Insufficient Corrosion by instrumentation Piston Improper seal rain or humidityDefective material Electronic control Diaphragm Insufficient Poor packaging failure shutdown devices controls and Inadequate Wheel protective devices Desiccant bolting Pneumatic control Material of omitted failure lades foil. root construction Fabrication Connected shroud Contamination wron Lubrication Corrosion and/or Welding error with dirt. etc Labyrinth erosion Foreign Improper heat Physical material left Thrust bearing Rapid wear treatment Insufficient oil General poor Pivoted pad bearing Fatigue Improper hardness Loading workmanship dam Roller/ball bearing Strength exceeded Wrong surface Water in oil finish T Cross-head piston Galling d Oil pump failure Imbalance Cylinder Wrong hardening Insufficient Low oil pressure method Lube passages not support Crankshaft PI Design for installation Assembly Improper filtration Unsatisfactory Improper fit piping support Contaminated oil proper Improper piping tolerances Craftsmanship flexibili after Parts omitted maintenance Undersized piping Parts in wrong Improper Inadequate tolerances Thefileisdownloadedfromwww.bzfxw.com
System design and specification responsibility Component manufacturer's responsibility Shipping and storage responsibility Installation responsibility Operations and maintenance responsibility Distress damage or failed components condition assumed (temperature, pressure, etc.) Incorrect physical property assumed (molecular weight, etc.) Specifications Inadequate lubrication system Insufficient control instrumentation Improper coupling Improper bearing Improper seal Insufficient shutdown devices Material of construction Corrosion and/or erosion Rapid wear Fatigue Strength exceeded Galling Wrong hardening method Design for installation Unsatisfactory piping support Improper piping flexibility Undersized piping Inadequate foundation Improper treatment Design Improper specification Wrong selection Design error Inadequate or wrong lubrication Inadequate liquid drain Critical speed Inadequate strength Inadequate controls and protective devices Fabrication Welding error Improper heat treatment Improper hardness Wrong surface finish Imbalance Lube passages not open Assembly Improper fit Improper tolerances Parts omitted Parts in wrong Parts/bolts not drainage Protective coating not applied Wrong coating used Equipment not cleaned Protection Insufficient protection Corrosion by salt Corrosion by rain or humidity Poor packaging Desiccant omitted Contamination with dirt, etc. Physical damage Loading damage Transport damage Insufficient support Unloading damage Cracking or separating Piping Misalignment Inadequate cleaning Inadequate support Assembly Misalignment Assembly damage Defective material Inadequate bolting Connected wrong Foreign material left in General poor workmanship Operating Slugs of liquid Process surging Control error Controls deactivated/not installed Operating error Auxiliaries Utility failure Insufficient instrumentation Electronic control failure Pneumatic control failure Lubrication Dirt in oil Insufficient oil Wrong lubricant Water in oil Oil pump failure Low oil pressure Plugged lines Improper filtration Contaminated oil Craftsmanship after maintenance Improper tolerances Sleeve bearing Seal Coupling Shaft Pinion/ball/turning gear Casing Rotor Impeller Shroud Piston Diaphragm Wheel Blades; foil, root, shroud Labyrinth Thrust bearing Pivoted pad bearing Roller/ball bearing Cross-head piston Cylinder Crankshaft The file is downloaded from www.bzfxw.com
System design Component Operations and Distress damage or and specification manufacturer's storage responsibility maintenance failed components responsibility responsibility responsibility responsibility Welding error Unsatisfactory soil data Poor alignment Improper surface Liquid ingestion Imbalance Improper fit Inadequate liquid Inadequat drain bearing contact General poor Design error Inadequate testing Assembly after Parts in wrong Parts omitted Misalignment r bolt ping stress Foreign material left in ng material of construction maintenance Schedule too long Requirements for Effective RCA. Performing an effective RCa requires an interdisciplinary approach in order to ensure that the results are correct and proper corrective actions are identified. In fact, most failures involve factors that spread across many disciplines such as metallurgy, mechanical engineering, hydraulics, electrical engineering, quality control operations, maintenance, human factors, and others. The analysis team on a complex failure will ideally represent a spectrum of expertise to ensure a very broad perspective The best analysis team leader must be a good communicator, have a broad background, be able to integrate factors, and be able to select the best expertise for the project. On less complex failures it is often beneficial to have an individual with a diverse background participate in addition to the specialists, once again to ensure a broader perspective. For example, a metallurgist may be more likely to report a metallurgical deficiency in a product that contributed to the failure,a fabricator is more likely to point to fabrication-related contributors, and a designer is more likely to identify design deficiencies. All of these may be important considerations, but one, all, or none may be a primary root cause. Problems related to people, procedures, environmental concerns, and other issues can also be treated effectively by conducting problem-solving processes and RCAs(although the main focus of this article and this volume is on materials failure analysIs)
System design and specification responsibility Component manufacturer's responsibility Shipping and storage responsibility Installation responsibility Operations and maintenance responsibility Distress damage or failed components Unsatisfactory soil data Liquid ingestion Inadequate liquid drain Design error tight Poor alignment Imbalance Inadequate bearing contact Inadequate testing Welding error Improper surface finish Improper fit General poor workmanship Assembly after maintenance Mechanical damage Parts in wrong Parts omitted Misalignment Improper bolting Imbalance Piping stress Foreign material left in Wrong material of construction Preventive maintenance Postponed Schedule too long Requirements for Effective RCA. Performing an effective RCA requires an interdisciplinary approach in order to ensure that the results are correct and proper corrective actions are identified. In fact, most failures involve factors that spread across many disciplines such as metallurgy, mechanical engineering, hydraulics, electrical engineering, quality control, operations, maintenance, human factors, and others. The analysis team on a complex failure will ideally represent a spectrum of expertise to ensure a very broad perspective. The best analysis team leader must be a good communicator, have a broad background, be able to integrate factors, and be able to select the best expertise for the project. On less complex failures it is often beneficial to have an individual with a diverse background participate in addition to the specialists, once again to ensure a broader perspective. For example, a metallurgist may be more likely to report a metallurgical deficiency in a product that contributed to the failure, a fabricator is more likely to point to fabrication-related contributors, and a designer is more likely to identify design deficiencies. All of these may be important considerations, but one, all, or none may be a primary root cause. Problems related to people, procedures, environmental concerns, and other issues can also be treated effectively by conducting problem-solving processes and RCAs (although the main focus of this article and this Volume is on materials failure analysis)