Look-Ahead Economic Dispatch With Adjustable Confidence Interval Based on a Truncated Versatile Distribution Model for Wind Power

Abstract

Wind power representation is an integral part of stochastic economic dispatch (ED). Compared with the deterministic approach, stochastic ED allocates system reserves more effectively. This improved reserve strategy heavily relies on the wind power distribution boundary area that is not consistent with classical unbounded distribution models. To address this issue, this paper proposes a new probability distribution model named truncated versatile distribution (TVD) for wind power, which is bounded and more accurate than classical models. The fitting effect and mathematical analysis of TVD are thoroughly discussed. Then, we propose a novel look-ahead dynamic ED method with an adjustable confidence interval (ACI) based on TVD, which not only enables system operators to maintain an acceptable predefined level of risk, but also seeks the optimal level of risk to balance the system reserve strategy and risk of power unbalance. With ACI, TVD offers a reliable wind power representation to search for the optimal level of risk in the dispatch model. In addition, the employment of TVD can simplify the wind power uncertainties by its analytical forms. The results show that compared with other classical distribution models, the novel distribution is a more accurate representation of wind power. Moreover, the ED model results in improved economic performances by ACI compared with traditional ones.