实时环境下动态无功优化建模研究

RESEARCH ON DYNAMIC MODELING FOR REACTIVE POWER OPTIMIZATION UNDER REAL-TIME CIRCUMSTANCE

为避免控制设备频繁操作,动态无功优化模型需引入变压器抽头和补偿装置投切开关的允许动作次数约束。一般做法是按全天24时段的预测负荷水平进行全局优化,寻优的难度大。作者将控制变量的动作次数约束还原为经济成本,并与当前时段下的电能损失费用共同构造目标函数,各时段的优化不再相关。采用灾变遗传算法便于求解,且无需依赖于负荷预测,适合于实时环境。文章用实例证明了该模型能恰当地描述负荷动态变化时的无功优化问题。

To avoid the frequent operation of control devices, the constraints of allowable operation times of tap-changing transformers and switches of reactive power compensators should be led into the dynamic reactive power optimization (DRPO) model. In general, the global optimization of reactive power is conducted by load levels forecasted in 24 time intervals in a day and the optimization is rather difficult. Here, the constraints of allowable operation times of the control variables are restored to economic costs which are used to construct the objective function with the cost for energy loss at the current time interval, consequently the optimizations in different time intervals are not interrelated each other. The test results show that the reactive power optimization problem can be appropriately described by the proposed model, when the loads are being changed dynamically, and is easy to solve by use of Cataclysmic Genetic Algorithm. This model is suitable to real-time condition and does not depend on the load-forecasted.