A probabilistic approach to monitoring the condition of complex technical equipment
Abstract
A probabilistic cascade-hierarchical approach is proposed for monitoring the condition of complex technical equipment in production, taking into account the types of controlled catches (software, hardware and mechanical) and types of control: technical means and the appropriate level of qualification of personnel based on polymodeling. In contrast to the deterministic control process, when sequential tests are carried out until a certain outcome is obtained, the proposed approach made it possible to form non-stationary rules for estimating the probabilities of process outcomes dynamically, relative to changing test conditions and the environment. The use of the Kullback-Leibler measure to assess the divergence of the probability distribution functions of defects in complex technical equipment, having high sensitivity, allowed us to form a scheme of a two-circuit decision support system for dynamic test planning of complex technical equipment, which ultimately increased the degree of validity of the decisions made.
