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 S5 > S7 > S4 > S6 > S11 > S3 > S10 > S9 > S13 > S2 > S8 > S12 > S1, where S1 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.

Online probabilistic operational safety assessment of multi-mode engineering systems using Bayesian methods

Online Probabilistic Assessment Of Operating Safety For Multi-Mode Engineering Systems Based On Bayesian Methods

Online Probabilistic Assessment of Operating Performance Based on Safety and Optimality Indices for Multimode Industrial Processes

A method of online safety assessment for industrial process operations based on Hopf bifurcation analysis

A multimode process monitoring method based Gaussian mixture models

Multi-mode process monitoring method based on PCA mixture model

Multimode Process Monitoring Based on Bayesian Method

GO-Bayes Method for System Modeling and Safety Analysis.

Multi-mode data-based joint fault detection method

A mixture of probabilistic predictable feature analysis for multi-mode dynamic process monitoring

Multimode Process Monitoring Approach Based on Moving Window Hidden Markov Model

Hierarchical Bayesian Operational Modal Analysis: Theory and Computations

A novel approach based on fault tree analysis and Bayesian network for multi-state reliability analysis of complex equipment systems

Towards Safe Multi-Task Bayesian Optimization

Bayesian Online Change Detection of Multimode Processes With Application to Wind Turbines

Statistical analysis and online monitoring for multimode processes with between-mode transitions

Modeling for evaluation of safety instrumented systems with heterogeneous components

Safety assessment for dynamic systems: A survey

Complex mechanical system safety prediction based on multidimensional indexes: An MBSA-PCA-BPNN method

A new perspective on Bayesian Operational Modal Analysis

A Markov chain Monte Carlo-based Bayesian framework for system identification and uncertainty estimation of full-scale structures