The impacts of correlation of multi-wind farms in day-ahead economic dispatching

In this paper, a two-dimensional versatile distribution was proposed to describe the joint cumulative probability density of relevant wind farms, which had a concise expression and was accurate in representing binary variables. Then it was applied to make sure that scheduling results could meet the reserve capacity requirement for a given confidence level. Based on this chance constrains model, four dispatched cases were worked out by considering the correlation of related wind farms and independent one. To optimize this nonlinear model, the Zoutendijk feasible direction method was introduced, which was efficient and time-saving. Eventually, a conclusion could be reached that by taking the correlation of related wind farms into consideration in dispatching, the reserve for wind power and the costs of wind power could be reduced, while unrelated farms should be treated independently for the sake of the grid's stabilization.

I. Introduction

As a widely spread green power, wind power resource shares a brilliant future under the background of lessening the usage of fossil fuel [1]. Nevertheless, the ratio of wind power keeps low in current power system [2], as its randomness and fluctuation are hard to overcome; the higher is the proportion of wind power is, the more spinning reserve capacity is required, which leads to more costs in generating. Besides increasing the costs, the stability of power grid was also faced with more uncertain factors. And their actual output values were more difficult to represent with multi-wind farms by traditional methods, which exerted an essential impact on balancing the active power [3]. So only by solving the problem of accurately describing the wind power and reducing the reserve capacity, can we increase the utilization of wind power.