Confidence capacity evaluation method and device based on vine copula and mixed offset normal distribution

Abstract

The invention discloses a confidence capacity evaluation method and device based on vine copula and mixed offset normal distribution. The method comprises the following steps: S1, inputting all new energy power generation output historical data in a system; s2, establishing a marginal distribution model of all new energy outputs by utilizing the offset normal distribution hybrid model; s3, establishing a correlation structure model among all new energy outputs by utilizing a vine copula method; s4, generating a plurality of new energy output scenes with correlation by utilizing the marginal distribution model and the correlation structure model of the new energy output; and S5, according to the generated new energy output scene, calculating the confidence capacity of the to-be-evaluated new energy power generation resource by utilizing a secant method. According to the method, the new energy power generation output historical data can be used to establish an accurate new energy outputmodel, a new energy output scene fitting the reality is generated, and the confidence capacity of the new energy power generation resources can be accurately evaluated.

Description

technical field [0001] The present application relates to the field of new energy, in particular to a new energy confidence capacity assessment method and device based on a mixture model of vine copula and offset normal distribution. Background technique [0002] The reason why new energy power generation resources cannot provide stable power generation capacity is the uncertainty of their output. This uncertainty not only comes from the volatility of the primary energy itself, but also is related to other primary energy sources and load correlations at different locations at the same time. [0003] The current evaluation algorithm for wind power confidence capacity is relatively mature, and the chord method is commonly used. Therefore, whether the confidence capacity assessment is accurate or not mainly depends on the accurate modeling of new energy uncertainty factors, and a large number of random scenarios using the established model. The modeling work is mainly divided ...

AI Technical Summary

Problems solved by technology

In previous research results, the mixed Gaussian model was often used to fit the irregular marginal distribution of new energy output. However, since the Gaussian distribution as a component cannot consider high-order statistics, more components are required to achieve a certain fitting accuracy. Adds complexity to parameter estimation

For the irregular correlation structure, the previous research results used the multidimensional elliptic copula function, but the multidimensional elliptic copula function cannot adapt to various atypical and asymmetric tail correlation structures, and the fitting effect is not ideal.

Images