Broadband non-Gaussian random extremum analysis and prediction method for monitoring data

Abstract

The invention provides a broadband non-Gaussian random extremum analysis and prediction method for monitoring data. Aiming at engineering monitoring data with relatively small sample size, such as load and structure response monitoring data, an autocorrelation function of a bottom-layer Gaussian random process of the monitoring data (non-Gaussian random process) is solved on the basis of a composite Hermite transformation model, so that a large number of bottom-layer Gaussian processes are obtained by a homologous multiplication algorithm; and a large number of non-Gaussian random processes are obtained through composite Hermite transformation. The non-Gaussian random process obtained by simulation has the same statistical characteristics and approximate random characteristics as the original female parent data, so that the reproduction data obtained by the method can be used for extreme value prediction. According to the method, a composite Hermite transformation method and a signal homologous multiplication method are organically combined, the method is applied to the field of extreme value prediction, the problem that a probability density model of a broadband non-Gaussian process cannot be constructed by a traditional probability method is solved, and the problem that large-capacity sample data are difficult to obtain by a traditional time domain method is solved.

Description

technical field [0001] The present invention relates to a monitoring data-oriented broadband non-Gaussian random extremum prediction method, which is based on compound Hermite transformation and signal homologous propagation, and relates to the technical field of random process extremum prediction and random signal reconstruction and propagation technology, especially civil engineering Broadband non-Gaussian stochastic extreme value prediction technology field in engineering and marine engineering. Background technique [0002] In the fields of civil engineering and ocean engineering, structures will always face challenges from complex environmental loads during their service. Whether it is the wind load and earthquake load that land structures often face, or the wave load and sea ice load that ocean structures often face, these loads are typical random processes. It is a complex physical process that random loads become random responses through structures. The random natu...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide a broadband non-Gaussian extreme value forecasting method for monitoring data, which is different from the traditional The probabilistic analysis method, by introducing the signal homologous propagation technology, avoids the problem that the probabilistic method cannot accurately interpret the broadband power spectrum, multi-modal power spectrum, and non-Gaussian response, and is different from the traditional time-domain analysis method, which does not require time-consuming Three-dimensional finite element calculation, but through the probability and statistics analysis, signal reconstruction and reproduction of a small amount of monitoring data, a large number of twin data simulations of monitoring data are realized, and extreme value forecasting is realized

Images