Interval model correction method taking correlation into account

Abstract

The invention discloses an interval model correction method taking the correlation into account. The method comprises the steps that firstly, correlative measurement responses are subjected to modeling by using an ellipsoidal convex model; secondly, through an interval optimizing method, reverse solution is conducted, and a corrected interval of model parameters is obtained; thirdly, an initial correlation matrix of the model parameters is set, and by combining the obtained model parameter interval, an ellipsoidal characteristic matrix of calculation responses is obtained by using an ellipsoidal characteristic matrix propagation formula; fourthly, the ellipsoidal characteristic matrix of the measurement responses and the ellipsoidal characteristic matrix of the calculation responses are subjected to least square optimization, and correction of the model parameter correlation is achieved. By means of the interval model correction method, the uncertainty interval of the model parameterscan be accurately corrected, moreover, the correction between the model parameters can be corrected, and therefore, the precision and the confidence level of the model are improved.

Description

technical field [0001] The invention relates to the technical field of model correction, in particular to an interval model correction method considering correlation. Background technique [0002] With the wide application of the finite element method, the finite element model correction technology based on the measured or experimental results has also been developed to a great extent. The traditional model correction methods assume that the parameters are deterministic. However, there are uncertainties in engineering practice. These uncertainties often exist in the material constitutive model parameters, structural geometric parameters, boundary conditions, initial conditions, Measurement information, cognitive judgment, etc. If deterministic methods and theories are still used to correct the finite element model, it will inevitably lead to unreliable correction results, which are quite different from the actual situation. Therefore, stochastic model correction methods co...

AI Technical Summary

Problems solved by technology

However, this method can only obtain the interval of model parameters, and cannot solve the correlation matrix between model parameters

[0005] It can be seen that most of the existing modal update methods use some algorithms to update the modal parameter range, and conduct research in terms of computational efficiency, without considering the correlation between model responses and the correlation between model parameters. Laws of Influence and Spread

Images