Failure Data

Failure analysis. Often, for a corrosion-related failure, data from the probes are examined to look for telltale signs that could have led to detection of the failure. In some cases, evidence can be found that process changes were occurring which led to the failure. This does not mean that the probes should have detected the failure itself. Determination of an imminent corrosion-related failure is not possible, even with the most advanced monitoring system. 

Table 1 shows prevalent examples of misconceptions about QRA. Many are actually generalizations that are too broadly applied. Two of the most common misconceptions concern (1) the lack of adequate equipment failure data and (2) the cost of performing QRA. 

QRA practitioners can use to satisfy some QRA objectives. Also, the American Institute of Chemical Engineers (AIChE) has sponsored a project to expand and improve the quality of component failure data for chemical industry use. And many process facilities have considerable equipment operating experience in maintenance files, operating logs, and the minds of operators and maintenance personnel. These data can be collected and combined with industrywide data to help achieve reasonable QRA objectives. However, care must be exercised to select data most representative of your specific system from the wide range available from various sources. Even data from your own plant may have to be modified (sometimes by a factor of 10 or more) to reflect your plant s current operating environment and maintenance practices. 

There are a variety of ways to express absolute QRA results. Absolute frequency results are estimates of the statistical likelihood of an accident occurring. Table 3 contains examples of typical statements of absolute frequency estimates. These estimates for complex system failures are usually synthesized using basic equipment failure and operator error data. Depending upon the availability, specificity, and quality of failure data, the estimates may have considerable statistical uncertainty (e.g., factors of 10 or more because of uncertainties in the input data alone). When reporting single-point estimates or best estimates of the expected frequency of rare events (i.e., events not expected to occur within the operating life of a plant), analysts sometimes provide a measure of the sensitivity of the results arising from data uncertainties. 

Synthesizing the frequencies of rare events involves (1) determining the important combinations of failures and circumstances that can cause the accidents of interest, (2) developing basic failure data from available... 

Once the variability risks, and q, have been calculated, the link with the particular failure mode(s) from an FMEA for each critical characteristic is made. However, determining this link, if not already evident, can be the most subjective part of the analysis and should ideally be a team-based activity. There may be many component characteristics and failure modes in a product and the matrix must be used to methodically work through this part of the analysis. Past failure data on similar products may be useful in this respect, highlighting those areas of the product that are most affected by variation. Variation in fit, performance or service life is of particular interest since controlling these kinds of variation is most closely allied with quality and reliability (Nelson, 1996). 

Support design FMEAs with failure data wherever possible Input from the eustomer and suppliers is important Should be reviewed at regular stages... 

Corley, J. E., Troubleshooting Turbomachinery Problems Using a Statistical Analysis of Failure Data, Proceedings of the 19th Turbomachinery Symposium, Texas A M University, College Station, TX, 1990, pp, 149-158. 

The objective is to estimate, numerically, the probability that a system composed of many components will fail. The obvious question is, "Why don t you just estimate the failure rate of the system from operating experience " There are three reasons IJ the system may not exist, so new data are not available, 2) the injuries and fatalities from the developmental learning experience are unacceptable - the risk must be known ahead of time, and 3) by designing redundancy, the probability of the system failing can be made acceptably remote in which case system failure data caimot be collected directly. The only practical way uses part failure statistics in a system model to estimate the system s reliability. 

OD -T alpha-lies, control, fault-tree description, failure data AND OR NOT K-of- N Bottom-up modularization and decompo-dtion of fault tree into t>est modular nepresentation Top event probability, time dependence, c s, r rCLs, and Linceitainty Option of not generating minimal cut >ets forquali ing fault ree IBM 360/370 Fortran IV, Available j from Software < enter... 

Reduce >f minimal cui failure data Mathematical combination of uncertainties output includes two moments of minimal cutsets and the lop event Johnson, empirical C le multiple. sy.siem fiiiJLuizjn with multiple data input descriptions can fit Johnsem-type distribution to the top event 1 t brnia... 

SPASM Fault tree or reduce tern equation, component failure data Combination (similar to BOUNDS) Lxignormal Works in conjunction with WAMCUT ... 

Recalculates sequence values after event failure data antPor cutsets have been modified. After choosing Sequence from the menu, the Sequence dialog appears listing all sequences in the current family. The pop-up menu lists Solve, Quantify, Uncertainty, CutSets. Display, and Time Dependent. 

Provides the means to recalculate end state values after modifying event failure data and/or cutsets. After choosing End State from the menu, the End State List appears listing all end states in... 

The NPRDS collects failure data on safety-related systems and components. At present, more than 60 plants are reporting data. The data are compiled and disseminated in periodic report,s to the participants of the program and other potential users. In addition, special searches of tl may be requested by the... 

The QRA was conducted by risk sts and design innel to determine the probability of explosive releases of the chemical. Fault tree analysis identified several combinations of equipment failures and operator errors that could cause the top event (reactor explosion), Failure data were obtained from plant ex ice and industry da%.ui,/uoes to quantify the fault trees to estimate the frequency of reactor explosions. The fault trees suggested several safety improv-... 

Figure 2-33 Hypothetical Two-Dimensional Failure Data and Design Curve...

Most comparisons of a failure criterion with failure data will be for the glass-epoxy data shown in Figure 2-36 as a function of off-axis angle 0 for both tension and compression loading [2-21]. The tension data are denoted by solid circles, and the compression data by solid squares. The tension data were obtained by use of dog-bone-shaped specimens, whereas the compression data were obtained by use of specimens with uniform rectangular cross sections. The shear strength for this glass-epoxy is 8 ksi (55 MPa) instead of the 6 ksi (41 MPa) in Table 2-3. 

Figure 2-36 Measured Failure Data for Glass-Epoxy (After Tsai [2-21])...

Other strength criteria are described by Sendeckyj [2-28]. Tennyson, MacDonald, and Nanyaro addressed the next logical step in a curve-fitting procedure, namely a third-order polynomial fit to failure data [2-29], However, the added complexity of their criterion has limited its use even though they identified some loading conditions under which their criterion is necessary to properly describe the actual failure behavior. 

When selecting in-process measures, try to use measures for which data are already available. For example, avoid using in-service failure data unless the maintenance systems can make this information readily available. These measures will be used to identify potential problems and correct them as early as possible. During the development of the integrated systems, data that will be available for in-process measurement should be identified and measures developed. These measures are most likely to be calculated annually as the volume of data required to provide useful data will be collected only over relatively long periods of time. 

Another example is a safety valve in standby service. If demands occur very infrequently, time-related stresses such as external corrosion may have a significant influence. Repeated demands in very dirty service could easily lead to faster degradation and failure, whereas repeated demands in lubricated service might actually enhance performance if the failure mode of interest is failure to open. Failure data based on time or demands can also be skewed if the relief valve is initially damaged or installed incorrectly. 

COMPL Data Bank for Component Failure Data... 

Mechanical, electrical, and electronic component failure data... 

The Department of Industrial Safety has been collecting and recording component failure data since 1978. For this purpose use is made of the following sources of information accessible (international) data-banks, literature and data... 

As failure data relating to mechanical components differ widely from source to source, TNO has set up a documentation system in which all relevant information is stored in one, uniform automated code called COMPI, which uses a component description code for the following information system of... 

The failure data relating to electronic and electric components are available in the form of handbooks. Failure rates are derived with the aid of calculation models based on statistical relations for which the incorporation of a (large) number of parameters is required. The following minimum of information is needed type of component, manufacturer and environmental factors. 

Pipeline Reliability An Investigation of Petroleum and Natural Data on frequency and cause of pipeline failures Data is specific to submarine and cross-country 49. 

This article presents an overview of the causes and frequency of failures for submarine and cross-country pipelines handling oil and natural gas. It gives several tables and charts which include information on the type of pipeline, the cause of the failure, and the number of failures. Data from failures in the US and the North Sea are included. Failure rates based on the total length of piping are calculated. 

---