作　　者： (胡伟）; (张玮灵）; (闵勇）; (陈磊）; (董昱）; (余锐）; (刘柏私）; (王磊）;

机构地区： 电力系统及发电设备控制和仿真国家重点实验室（清华大学电机系）,北京市海淀区100084

出　　处： 《中国电机工程学报》 2017年第16期4567-4576,共10页

摘　　要： 电力系统紧急控制是故障消除后保持系统稳定性、防止事故扩大的重要手段。相比于目前电网普遍使用的离线决策,实时紧急控制决策能够提供更准确有效的控制措施。但实时决策对计算速度的要求极高,传统方法无法胜任。基于支持向量机技术,该文研究了电力系统紧急控制实时决策方法,关键在于利用支持向量机从大量的仿真数据中挖掘系统稳定规则,并用于构造可以实时求取的稳定性约束,由此将大量的仿真搜索从实时阶段转移到离线阶段。该文首先建立紧急控制决策的优化模型;然后利用灵敏度方法将模型中的稳定性约束线性化;接着通过改造支持向量机的分类表达式将其用于拟合,提出能够实时求取的稳定裕度指标,实现灵敏度求解;最后在IEEE39节点系统中进行算例仿真,仿真结果显示:所提出的稳定性指标能够准确反映系统的稳定裕度,同时针对不同的失稳状况该文提出的决策方法能有效恢复系统稳定性。 Emergency control of power system is an important means to keep system stability and prevent accidents expanding after the failure is eliminated. Compared with the off-line decision-making widely used in the current power grid, the real-time emergency control decision can provide more accurate and effective control measures. But the real-time decision is very demanding for computing speed, and the traditional method is not competent. Based on support vector machines （SVM）, a real-time decision making method for emergency control of power system has been studied. The key technology was the use of SVM to dig system stability rules from a large amount of simulation data, so as to obtain the stability constraints in the real-time structure, and thus a large number of simulation searches can shift from the real-time phase to the offline stage. This paper firstly established the optimization model of the emergency control strategy, and the stability constraints in the model were linearized by the sensitivity method. By transforming the classification expression of SVM, the stability margin index can be obtained in real-time, and the sensitivity was solved. Finally, the simulation results in IEEE-39 buses system showed that, the stability index proposed can accurately reflect the stability margin of the system; at the same time the decision method of the proposed stability condition can effectively restore the stability of the system according to different disturbances.

关 键 词： 电力系统 紧急控制 实时决策 支持向量机