Complex mechanical system safety prediction based on multidimensional indexes: An MBSA-PCA-BPNN method
ENGINEERING FAILURE ANALYSIS(2024)
Abstract
Conventional methods are progressively insufficient to predict the safety of overall systems characterized by complex subsystem coupling, resulting from empirical and qualitative. Therefore, an innovative analysis that could quantify system safety with a standardized analysis accounting for the influence of multidimensional indexes that characterize different subsystems is proposed in this study. Specifically, an aviation piston engine is analyzed in this study as an illustration of the complex system. Firstly, a nominal model considering multiple subsystems of aviation piston engines is developed. Then, the indexes adopted to characterize the safety of the overall engine are obtained by injecting possible faults into the nominal model. Finally, a PCA-BPNN model is constructed to downscale the indexes and predict the safety rating. The application of the method to a total of 13 states containing normal and fault in the engine shows that the states ranking based on different indexes is varied as influenced by coupling. However, the ranking of states is consistent by PCA, proving the synthesis and rationalization of the analysis based on a multidimensional perspective. Specifically, the ranking of states from highest to lowest of severity is S-5 > S-7 > S-4 > S-6 > S-11 > S-3 > S-10 > S-9 > S-13 > S-2 > S-8 > S-12 > S-1, where S-1 represents the normal state. Moreover, the predictive goodness of fit of the safety prediction model PCA-BPNN incorporating severity and probability achieves 0.986, which not only improves the efficiency but also avoids the impact of multiple covariances on the model performance. Results of the quantitative classification of safety rating for the case show that the inlet valve opening angle parameter has the highest safety rating 11, followed by the exhaust valve opening angle with safety rating 8. Therefore, a detailed safety analysis and design of the valve opening angle parameter is required, especially for the intake valve. Generally, the proposed method could provide additional clarification to the traditional local perspective analysis by considering the coupling, thus guiding the improvement from a safety aspect.
MoreTranslated text
Key words
Safety prediction,Aviation piston engine,Complex mechanical system,Model based system safety analysis,Multidimensional index,PCA,BPNN
AI Read Science
Must-Reading Tree
Example
![](https://originalfileserver.aminer.cn/sys/aminer/pubs/mrt_preview.jpeg)
Generate MRT to find the research sequence of this paper
Chat Paper
Summary is being generated by the instructions you defined