A framework for the reliability characteristics estimation on the experimental data

This paper presents a framework for the reliability characteristics estimation on the experimental data. The framework consists of the set of failures, the experimental data sources, the set of reliability characteristics, the estimation methods, the estimation goodness tests and the expert system supporting estimation and statistical inference processes. The experimental data was obtained with two main theoretical means: control theory and diagnostics. It has been observed that changes of the functioning condition potential could serve to determine the symptoms of transient (momentary) failures and the changes in the potential of the technical condition to determine the symptoms of parametric faults. Such an information is sufficient to calculate reliability characteristics before dangerous catastrophic failures occur and to calculate reliability characteristics for single object without having to deal with a numerous set of objects. The set of reliability characteristics models consists of the analysed failures combination and distribution functions used in the reliability engineering. The set of estimation methods includes analytical and numerical ones. The proposed expert system supporting estimation and statistical inference processes consists of data base (the experimental data), rule base and inferring machine.