The Approach And Tool For System-level Software FMEA

In recent years, computer software has played an indispensable part in all kinds of fields. However, computer software may makes some failure like other products as well. When it occurs, not only will it lead to the task failure, but also result in some catastrophic consequeces. In order to improve the qualities of computer software, doing some reliability analyses during the software developing process is necessary and its improtance is self-evidence. Among the numerous methods of reliability analysis, software FMEA has proved to be one of the most effective methods to improve the qualities of computer software. Software FMEA can be divided into system-level FMEA and detailed-level FMEA. This paper mainly focuses on the system-level software FMEA.The traditional method of system-level software FMEA ,which mainly depends on the experiece of analyzer,also has some shortages, such as low accuracy, lacking of objectivity ,automated analysis tools and so on. Therefore, in this paper, we proposes an approach to construct both software hierarchical structure and dependece relation models , at the same time,it designs a supplementary analysis tool which can be used to support the target system-level FMEA.First of all, the thesis proposes an approach for constructing software hierarchical structure and dependence relation model, which can divide the target software into system level, describe the logical relationship between modules and track,the underlying software modules to determine failure of the impact of high-level modules.We proposed an approach of failure effect analysis to find the accessible nodes of graph, which provides a strong basis for accurate and objective system-level software FMEA. Secondly, through collecting a large number of software failure cases, this thesis concludes "Classification Lists of General Software Failure Modes", "Classification Lists of Embedded Software-specific Failure Modes" and "The Reasons for Software Failure". It also summarizes the extraction principles of software failure modes and failure analysis for different software, which accumulates valuable experience for the future system-level software FMEA.At the same time, in order to adapt to the development of internet and meet the needs of large-scale software analysis, this thesis designs and implements the assisted analysis tools of system-level software FMEA. This tool can be used by software developers in the design process. It can also improve the automation of system-level software FMEA. We standardize the design of the underlying database with the principles of database design, which makes it not only be able to store the data of system-level software FMEA generation, but also be able to effectively reduce data redundancy and ensure the efficient operation of the system. Finally, we select a typical case, and then conduct the whole system-level software FMEA just as described above. All this proved the feasibility and effectiveness of the method.