AUTOMATION OF FAILURE MODES, EFFECTS, AND DIAGNOSTIC ANALYSIS REPORTS GENERATION Alexander Yasko*, student of department 503, group 535a; Eugene Babeshko, senior lecturer of department 503 National Aerospace University named after N.E. Zhukovsky "KhAI"
The reliability and safety of modern computer systems are the very first features to think of when the products for critical domains are being developed. The FMEDA (Failure Modes, Effects, and Diagnostic Analysis) is one of the most preferred techniques recommended by different experts to comply with reliability and safety requirements from different standards in this field. Usually the FMEDA process is easy enough to perform, but the difficulty is increasing with the complexity of the product. The experts have to work with a great amount of datasheets, non-standardized databases to find characteristics (such as failure modes, distribution ratios) for each electrical component of the system, as well as analyze huge number of circuit diagrams to identify the most critical places. Thus, the FMEDA technique needs to be automated to facilitate the overall reliability and safety analysis process and make it less dependent from experts’ decisions. During the research, it was identified that companies that provide Assessments & Certification services have their own proprietary automation tools. Examples of such companies include exida and ENCO. They offer automation of FMEDA, but experts still have to do routine work using not uniformed sources of electronic parts. In this report, the general approach of FMEDA technique is reviewed, existing solutions are analyzed, experience of different companies on FMEDA related activities during certification is analyzed, in particular RPC Radiy during SIL3 licensing of the RadICS platform (set of the FPGA-based modules). The means of automation of the FMEDA are considered. The tool for facilitating the FMEDA reports generation is being designed. While performing the assessment of electric circuits, it allows to assign easily the characteristics, failure rates, failure modes, other relevant information to the system components. The dependence of experts’ decisions is to be decreased using the unified failure database that is to be developed. It is considered to contain the characteristics, failure specifications of electrical components (such as resistors, capacitors, transistors etc.) and it is supposed to be connected for the tool being designed. The next steps of the FMEDA automation include the development of the tool, its introduction into the assessment and analysis techniques. Furthermore, the unified database of components characteristics and faults is to be developed. *Scientific supervisor – Doctor of Science, Professor Vyacheslav Kharchenko.