Wind turbine generator parameter identification method based on Bayesian neural network

Abstract

The invention discloses a wind turbine generator parameter identification method based on a Bayesian neural network, and the method comprises the following steps: S1, collecting historical data of a wind turbine generator, and initializing Bayesian neural network model parameters; S2, dividing historical data of all wind turbine generators into training data and test data; S3, calculating networkoutput by using the training data; S4, updating the weight of the Bayesian neural network model; and S5, calculating a global error, judging whether the requirement is met or not, if so, obtaining a final network weight matrix, and ending the learning algorithm, otherwise, returning to S3, and entering the next round of learning; and S6, calculating network output by using the test data and the network weight to obtain the parameter identification result of the wind turbine generator. According to the method, the Bayesian theory and the neural network model are combined, compared with a traditional parameter identification method, the method considers the influence of the uncertainty change of the external environment in the identification process, and the method has the advantages that the global error is easy to converge, and the number of iteration steps is small.

Description

Technical field [0001] The present invention relates to the field of smart grid power distribution, and more specifically, to a method for identifying parameters of wind turbines based on Bayesian neural network. Background technique [0002] As new energy sources such as wind power are connected to the power system, the uncertainty that exists becomes more and more obvious. The existence of uncertainty makes the system model difficult to accurately model with fixed parameters. Ignoring these model errors cannot obtain calculation results consistent with the actual power grid, and cannot accurately determine the stable status of the power system. [0003] Traditional model parameter identification methods mainly include least square method, gradient descent method and neural network method. The first two identification methods are mainly used for the identification of linear model parameters, while the neural network method is mainly used for the identification of nonlinear model...

AI Technical Summary

Problems solved by technology

However, due to the uncertainty of external conditions such as wind speed during wind power grid-connected operation, some parameters of wind turbines will continue to change, and traditional deterministic parameter identification methods cannot cope with this random change.

Images