Portal de Periódicos da CAPES

Method and algorithm for preventing extreme operating conditions of a complex technical object

2021; American Institute of Physics; Volume: 2402; Linguagem: Inglês

10.1063/5.0073463

1935-0465

Oleg V. Prokofiev, Mihail Yu. Miheev, Alexander E. Savochkin,

Advanced Statistical Methods and Models

The technical object operation can change significantly and approach the extreme value due to the occurrence of events, which are hard to predict. The structure change in the time series of a controlled value may indicate a critical event. Decision-making for the prevention of pre-emergency situations is based on the short time series monitoring, as well as the time series structural stability analysis. The correct operation of a complex technical object implies the absence of trend breaks in time series generated from the readings of parametric instrument sensors. The proposed Chow test modification allows finding possible structural instabilities in the entire series without the participation of an expert. The modified Chow test allows finding the changes in the trend slope. The proposed solution does not require expert estimates, allows using short sections of time series, and requires a small number of arithmetic operations. The proposed tools can be used in engineering applications. The conducted research leads to the following conclusions — the modified break identification algorithm adapted for automatic use without expert estimates increases the accuracy of the conclusions about the short time series structure changes.