This paper focuses on the interchange and generation scheduling problem of multi-area power systems, where decentralized decision procedure is preferred and the uncertainties of nodal injections should be considered. The problem is formulated as a multi-area robust security constrained unit commitment model, with a novel uncertainty set specified by the variance of system net load. We prove that such an uncertainty set can provide a more precise probabilistic guarantee than conventional budgeted sets. To solve the problem in decentralized manner, an upper-level robust convex optimization model is proposed to optimize the tie-line power flow and derive a generation interval for each area. Then through the coordinated uncertainty requirement decided by the generation interval, each regional system operator can solve a decoupled robust unit commitment problem independently. Besides, a modified outer approximation algorithm is developed to obtain higher-quality solution of the bilinear programming involved in the process. Simulation results on a two-area system demonstrate the effectiveness of the uncertainty set, and show the economic efficiencies of unit commitment solutions acquired from the hierarchical method.

中文翻译：

---

具有新不确定性集的多区域电力系统鲁棒 SCUC 的分层方法

本文重点研究多区域电力系统的交换和发电调度问题，其中优选分散决策程序并应考虑节点注入的不确定性。该问题被表述为多区域鲁棒安全约束单元承诺模型，具有由系统净负载方差指定的新不确定性集。我们证明，这样的不确定集可以提供比传统预算集更精确的概率保证。为解决分散方式的问题，提出了一种上层鲁棒凸优化模型来优化联络线潮流并推导出每个区域的发电间隔。然后通过生成间隔决定的协调不确定性要求，每个区域系统运营商都可以独立解决一个解耦的鲁棒机组承诺问题。此外，还开发了一种改进的外逼近算法，以获得该过程中涉及的双线性规划的更高质量的解决方案。两区域系统的仿真结果证明了不确定性集的有效性，并显示了从分层方法获得的单位承诺解决方案的经济效率。