Reliability Analysis

Reliability according to DIN 40041 is defined as the capability of an observation entity (components, systems, plants, process) to meet expected performance criteria during a given time period. Its analysis is done using stochastical methods, in particular distribution functions for the failure event.

Your benefits:

Reliability analyses provide objective and quantifiable quality assessments of your technology. The real operational behaviour of your technical product or service will become more transparent. Last but not least you will gain additional selling or buying points..

Important reliability parameters are mean value, variance, variation coefficient, median and other quantiles, confidence intervals as well as graphs of distribution function, reliability function, density and failure rate including their confidence intervals.

The above parameters are also the basis for further product assessments:

• Detection of weaknesses and cost drivers
• Verification of requirements for tenders
• Forecasts of reliability behaviour of new products
• reliability assessments of larger systems (FMEA and FTA)
• Approximate calculation of remaining lifetime
• Verification of reliability characteristics of the products
• Verification of required number of spare part units
• Calculation of life cycle costs (LCC)
• Efficiency investigations and process optimisation in maintenance

Reliability must be substantiated by secure information. The samples which are needed to this end can be generated almost automatically by applying our analytical software to the databases of your IT system.

Reliability changes with the development of your product. We will assist you with the following:

• a standardised documentation of your examination results
• the setting up of a parameter directory as scientific baseline for reliability forecasts

Approach:

• Sampling, i.e. listing of all down-times and up-times or other censoring
• Computation of the empirical distribution function or individual empirical characteristic values such as expectancy value and other things.
• Statistic modelling of the failure development (for example Weibullanalysis)
• Computation of further characteristics (failure rate, confidence intervals,…)
• Interpretation of the results, if necessary data correction and new analysis
• Processing of the results

Selected References:

• "BestVal" project and further analyses for vehicle series, Deutsche Bahn AG, produkt range vehicle TFE 5
• "Natur" project (... New AspecTs of reliabilty examinations for DRAM memory chips) by Infineon AG
• "FXM" project (Field eXperience Management)by SIEMENS AG