STATISTICAL METHODS FOR ASSESSING THE STABILITY OF FUNCTIONING OF TECHNICAL SYSTEMS 

Klyachkin Vladimir Nikolaevich, doctor of technical sciences, professor, sub-department of applied mathematics and informatics, Ulyanovsk State Technical University (432027, 32 Severny Venec street, Ulyanovsk, Russia), v_kl@mail.ru
Karpunina Irina Nikolaevna, candidate of technical sciences, associate professor, sub-department of general professional disciplines, Ulyanovsk Institute of Civil Aviation (432071, 8/8 Mozhaiskiy street, Ulyanovsk, Russia), karpunina53l@yandex.ru.

519.248: 681.518.5

10.21685/2307-4205-2018-2-5

Background. Stability of functioning of the system often determines its performance. Instability can lead to system or stop the occurrence of an emergency situation. An important task is the early detection of violations of process stability operations.
Materials and methods. Violations are manifested in the form of statistical characteristics, therefore, to find them can be used methods and algorithms for statistical process control. Such approach allows to reveal the violation occurred before exit controlled performance limits. Thus, as a rule, simultaneously monitoring and mid-level process and its dispersion. Selected statistical tools for control: for independent indicators used standard Shewhart’s charts, for correlated methods and algorithms, based on the use of multivariate Hotelling’s statistics and generalized variance.
Results. This study allowed us to propose the following methodology for assessing the stability of technical systems: a study of working conditions analogical systems and identification of possible disturbances; calculation of basic statistical characteristics of the process; selection of statistical tools for follow-up depending on the alleged violations and the level of correlativeness of parameters; provodenie for continuous monitoring of the functioning of the system for diagnostics of stability.
Conclusions. The proposed methodology would detect violation of process stability functioning of the technical object and prevent the possible emergency

technical diagnostics, detection, Shewhart’s control chart, Hotelling’s algorithm, generalized variance