The software support for an identification of stochastically loaded parts of mechanically structures

Abstract

The paper deals with one of possible ways of an identification of stochastically loaded mechanically structures. The purpose of this approach is to find an algorithm of a forecasting control of their working in real working conditions. It deals with a proposal of an application of vector time series moving average models (VARMA). Their parameters are possible to determine using the nonlinear modification of the least squares method. The paper contains a theoretical principle of problems solved and a description of a real testing methodise. There is added a brief interpretation of results of used software verification.