ADAPTIVE RELIABILITY MANAGEMENT OF TECHNICAL SYSTEMS BASED ON BAYESIAN NETWORKS FOR ROOT CAUSE FAILURE DETECTION

Le résumé fourni par la source

The relevance of the study is determined by the increasing complexity of technical systems, the intensification of their operation, and the growing requirements for reliability under conditions of uncertainty and variability of operating modes.Traditional diagnostic approaches do not ensure effective detection of latent defects and causal relationships between failures, which necessitates the development and implementation of adaptive reliability management methods.The aim of the study is to improve the effectiveness of reliability management of technical systems through the application of Bayesian networks for adaptive identification of root causes of failures.The research employs methods of system analysis, generalization, comparison, and structural-functional modeling to investigate reliability management approaches, the principles of Bayesian network construction, and mechanisms for integrating diagnostic and operational data.The results of the study demonstrate that modern approaches to reliability management are evolving toward integrated probabilistic and adaptive models.The capabilities of Bayesian networks for modeling causal relationships and diagnosing technical conditions have been examined.The principles of constructing adaptive models based on the integration of heterogeneous data have been generalized.The main scientific and practical problems have been identified, including the complexity of model structure formation, limited and incomplete data, non-stationary operating conditions, and computational constraints.It has been proven that the efficiency of reliability management increases through the use of integrated models capable of adaptation and continuous updating.The conclusions confirm that the application of Bayesian networks combined with the integration of diagnostic and operational data improves the accuracy of root cause failure detection and enhances the validity of decision-making.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.