Error Analysis of Ultra Short Term Wind Power Prediction Model and Effect on the Power System Frequency

Abstract

Objective: the forecasting error analysis of wind power is the preparation of power system optimization scheduling for large scale wind power access, which can improve the security and economy of the system operation. Methods: in this paper, we use a piecewise exponential distribution model to predict the ultra short term wind power error and then estimate the parameters. Process: the case we used is from Northern Ireland, we forecast the probability and precision of wind power on the basis of Normal distribution model, Laplace distribution model, Cauchy distribution model, Beta distribution model and the proposed piecewise exponential distribution model, and study its effect on the frequency of power system. Conclusion: the prediction error distribution model of the sub index wind power forecasting error can be used to mine the relative information of the actual error distribution, in addition, it’s convenient to implement and easy to be used in calculus, it can be applied to describe the error distribution of the multiple time scale prediction, so it has more advantages in the error analysis.