Failure mode and effects analysis

FMEA (english Failure Mode and Effects Analysis, dt " Failure Mode and Effect Analysis " or simply " impact analysis ") and FMECA ( engl. Failure Mode and Effects and Criticality Analysis) are analytical methods of reliability engineering, to find potential vulnerabilities. In the context of quality management and safety management, the FMEA is used to avoid errors and increase the technical reliability preventively. The FMEA is particularly applied in the design and development phase of new products or processes, but demanded by suppliers of standard parts for the automotive manufacturers also other industries.

Base

FMEA follows the basic idea of a precautionary error prevention rather than a remedial error detection and correction (fault management) through early identification and assessment of potential causes of errors already in the design phase. Thus resulting control and error correction costs are otherwise in the production phase or even in the field avoided ( the customer ) and overall costs reduced. Through a systematic approach and the information obtained thereby also the repetition of design defects in new products and processes is avoided.

The methodology of FMEA is to be applied already in early stages of product development (planning and development ) within the product life cycle, as a Kosten-/Nutzenoptimierung in the development phase is the most economical (preventive error prevention ). The later an error is discovered, the more difficult and expensive its correction will be.

Species

FMEA can be divided in several ways:

The system FMEA and the HW / SW / design FMEA for so-called product - FMEA was summarized by the VDA 2007, as the considered system usually can not be clearly resolved.

The system FMEA is applied in the development process. Your task is on the one hand, to investigate the product on meeting the requirements set out in the requirement specification 's functions, on the other hand, especially failure modes that lead to non-compliance is to collect and evaluate. In this case, appropriate measures to prevent or discovery of potential errors to be planned for all risk-related parts of a product. The system FMEA at component level corresponds to the previous definition of the Design FMEA. It is used for analysis of all component features that are necessary for the fulfillment of the required component function.

The system FMEA process is still used within the production planning process. It builds logically on the results of the subordinate FMEA. An error in the system FMEA, the cause lies in the manufacturing process shall be incorporated as a logical error in the Process FMEA. The object of the System FMEA process is to examine the entire manufacturing process of a product on the suitability for manufacture of the product back. Here, for any errors that may occur during the manufacture of the product, to plan appropriate measures to avoid or discovery.

Application

In the first application, a team of employees from different business functions ( interdisciplinary team ) is formed. The assessment shall consider in particular the design, development, testing, production planning, manufacturing execution, quality management, etc. The analysis process itself is then carried out with use of standard forms (QS -9000 ) or equivalent software in a formalized manner (VDA 4.2).

Contains FMEA

• A limitation of the system under consideration,
• Structuring of the system under consideration,
• Definitions of functions of the structural elements,
• An analysis of potential causes of errors, error types and error sequences, either directly ( eg, using the W - questions ) are derived from the functions of the structural elements,
• A risk assessment,
• Measures or solutions to prioritized risks
• A pursuit of agreed measures and avoidance and discovery
• A residual risk assessment or evaluation.

Potential errors are analyzed by the fault is located, determines the type of fault, describes the error sequence, and then the cause is determined. For the determination of possible error causes a so-called cause-and- effect diagram is often created. It is possible that even direct evidence of possible measures can be derived to avoid errors due to a detected error.

The ratios B, A and E on the importance (. The error sequence, English Severity S), occurrence probability (the error cause, English Occurrence O. ) And detection probability ( the error or its cause; possibly also the sequence;. Engl Detection D) a basis for risk assessment. The key figures are integer values ​​between 1 and 10 and will be awarded with the aid of Review catalogs.

By calculating the risk priority number (RPN ), an attempt is made ​​to create a ranking of risks. The RPZ produced by multiplying the B-, A- and E - valuation numbers () and can therefore take values ​​between 1 and 1000. There is the claim that the RPZ, / allowed at least in comparison with other RPZ the same FMEA, a statement within the meaning better worse.

The aim of the RPZ to estimate the importance and rank of error in order to derive priorities for action to be taken is repeatedly asked in question. Attempts are being made to the characteristic () in addition to or alternatively to operate. In DRBFM, the FMEA methodology used in Toyota, remains under the definition of key figures in its entirety. Measures are there exclusively determined by common sense or as a result of team discussion.

Measures

A distinction is made between preventive measures and detection actions. Avoidance measures serve to justify the rating number for the probability of occurrence (A) and detection actions to justify the rating number for the probability of detection (E).

Measures the actual state (initial state ) are usually documented in charge and without appointment. With the valuation numbers for B, A and E, the RPN is calculated for the actual situation.

Additional measures are aimed at

• To reduce the probability of occurrence of an error ( for example, by the installation of improved components).
• To increase the likelihood of potential discover errors cause by, for example, additional tests are provided.

Any modification to the product or process, a new risk assessment is required for the affected by the change area. An " improved component ", for example, in addition to positive aspects have a higher weight or a higher power consumption.

The discovery measure is usually not discover for pragmatic reasons, the cause of the error, but the error or the error sequence. The documentation is still mostly in the cause of the error.

If a continuous improvement process are supported, then the measures in measures objects, each with its own A - and E - score value and thus of his own RPZ be grouped.

In the automotive industry, the measures stands are also categorized to respond in case of error causes on the topics of development, and field service in design FMEA. Each category has its own RPN.

The risk assessment takes place in the current FMEA not alone anymore by the already mentioned RPZ place, but rather in the following sequence:

Highest priorities have high meanings (10 ), then the product is of importance and probability of occurrence considered (B * A), this is also referred to as criticality or Technical risk (Considerations underlying the valuation numbers stored catalogs, A = x is a region and no fixed ppm number for the probability of occurrence of the error ).

Only then reaches for the prioritization of the remaining points the RPZ.

Assessment

The assessment is done by interdisciplinary teams to "1" assigned points for each of "10". There will always be graded from the higher valuation for the lower rating.

• Importance or severity of the error sequence is evaluated from the perspective of the customer ( high = " 10" to low = " 1").
• Probability of occurrence of the cause ( high = " 10" to low = " 1")
• Discover probability of the cause or the error in the process, before delivery to the customer ( low = " 10" to high = "1")

The customer can account both the end user as well (for example, in-house ) to be intermediate customer who demands the FMEA. Risk priority numbers ( RPN) can be used to rank the agreement of countermeasures in the development process. The RPN alone is not suitable for the assessment of potential risks. An example of RPN 120 may have arisen in a variety of ways, such as of W x O x D = 10 x 3 x 4 or from 5 x 8 x 3 one meaning of B = 10 and a rather moderate discovery of E = 4 is less acceptable as an error sequence, evaluated with B = 5, the occurs very frequently (A = 8), but will be well detected ( e = 3).

An assessment of the absolute risk potential can be derived from the factors the significance and likelihood of occurrence.

Different companies have defined your own, ajar to the standard works, catalogs risk evaluation and criteria for the adoption of measures to reduce risk.

After initial and processed measures, a renewed risk assessment: It is checked by repeated determination of a Risk Priority Number RPN (or RPN ), whether the envisaged measures promise a satisfactory result ( the meaning of the error sequence will remain unchanged ). If the result is not the required quality requirements of the customer, other preventive or discovery approach should be taken and / or solutions to be developed.

The VDA Volumes 4, Part 2 and 3 suggest details a systematic approach.

Historical

For the first time a description of the FMEA method was published as United States Military Procedure MIL- P- 1629 - Procedures for Performing a Failure Mode, Effects and Criticality Analysis; November 9, 1949. Widespread use of FMEA in the field of automotive industry was initiated by Ford after it was sensational problems model Ford Pinto in the 70s.

In the 1970s, the three largest U.S. car company General Motors, Ford and Chrysler faced its suppliers, each with individual FMEA guidelines. One of them had, for example, only five rating values ​​for B, A and E. On the initiative of suppliers is a unification in the form of the QS -9000 FMEA reached at the beginning of the 1980s; the Big Three (Ford, GM and Chrysler) took it the FMEA method description of Ford as a basis and added a few essential supplements in the form of attachments as appropriate, eg their own symbols for classification. In the following years was the widespread introduction of FMEA in the supplier industry. By the Association of the Automotive Industry ( VDA) 1996 was published an improved FMEA methodology. In the available since 2002 third edition of the QS -9000 FMEA methods description, some elements of the VDA approach were adopted. 1997 Toyota for the first time described a change -focused FMEA, which is known as DRBFM methodology today. In March 2007, the VDA FMEA 2nd edition went to press. After the QS -9000 standards are no longer current AIAG as editor of the now available AIAG FMEA 4th edition (June 2008) has been selected.

The fields of application of FMEA have widened over time. Originally located in the military area has gained recognition via an intermediate stage " Aerospace " in the automotive sector, the FMEA. Since the FMEA is a universal model based methods, they are also found in other areas in which work systematically, their application fields, such as medical, food industry ( as HACCP ) system engineering, software development.

Norms and Standards

For the FMEA there are various standards depending on the application context. A general context- unspecific Standardization was carried out in 1980 by the DIN 25448 with the keyword " failure modes and effects analysis ." ' This standard was updated in 2006 by the DIN EN 60812 with the keyword " failure mode and effects analysis ". There are also numerous context-specific standardization, below is a small selection: