The calculating PHM cluster: CH&P mathematical models and algorithms of early prognosis of failure

Abstract

This work describes mathematical models and computing cluster for early failure prognosis and accurate estimates of remaining useful life (RUL) for technical objects: internal combustion engines, gas turbine, hydroelectric turbines, wind turbines, etc. The hierarchy of mathematical models for prognosis (CH&P) is based on a hierarchy of degrees of developed failure, and solves the problem of accurate assessment of RUL; determines the required physical parameters for the prediction and risk assessment; classifies the signs and their evolution at all stages of development. In the absence of early incipient fault the mathematical model identifies incipient of fault cause, the time evolution of which leads to the appearance of early incipient fault. In the absence of incipient of fault cause the hierarchical mathematical model analyzes the state of the system using the methods of symbolic and topological dynamics to identify the evolution of symbolic hidden trajectories of the observed signals, which leads to Incipient of hidden fault cause. Thus, the hierarchical mathematical model provides the earliest prognosis of occurrence of failure causes. It is also noted that in the analysis stage of hidden trajectories (preventive prognosis) is possible a physical reversibility in the technical system. There is a legitimate question about the implementation of the automatic stochastic management by system in real time in order to avoid failure at the stage of the appearance of their hidden causes.