Application of E-Regions of the Parameters of the Model Function of the Signal for the Extraordinary Situations Detection

Abstract

Methods of solution of a problem of an estimation of uncompensated errors of the measurements caused by presence of noises and interferences, that not submitting to statistical regularities are investigated. The model of the signal presented in the form of a composition multiparameter quasidetermined model function and fixed on interval of measurement a background function is considered. The background function is described by an arbitrary ensemble of functions whose range of variation is limited by the E-layer. The proposed model provides the calculation of interval estimates of signal parameters for confidence values close to unity. The technique of calculation of interval errors of parameters on the basis of model of E-areas and nonlinear model function is described. This technique allows calculating interval estimations of parameters for a particular of a signal realization, provided that the range of values of the background function is limited. It is shown that defined in the space of parameters E-regions characterize the state of the control object. Based on the developed algorithm for diagnosing the state of an object from the particular signal realisation using E-regions in the space of parameters of a model function, an approach to identifying abnormal situations is proposed. Based on the developed algorithm for diagnosing the state of an object from the implementation of a signal using E-regions in the space of parameters of a model function, an approach to identifying abnormal situations is proposed. The results of practical application of E-regions for monitoring the process of power consumption of electric energy and other energy resources are presented.