考虑离散马尔可夫模型的多工况电力系统自适应控制策略  

Adaptive Control Strategy Considering Discrete Markov Model for Power System With Multiple Operating Conditions

作　　者：马静 高翔[2] 李益楠 刘畅 王增平 

机构地区：[1]新能源电力系统国家重点实验室(华北电力大学),北京市昌平区102206 [2]国网冀北电力有限公司廊坊供电公司,河北省廊坊市065000

出　　处：《电网技术》2015年第10期2856-2861,共6页Power System Technology

基　　金：国家自然科学基金项目(51277193);中央高校基本科研业务费专项资金项目(2014ZZD02)~~

摘　　要：针对连锁故障导致电力系统运行工况变化较大的问题,提出一种考虑离散马尔可夫模型(Markov)的多工况电力系统自适应控制策略。首先,建立考虑电力系统多工况因素的离散Markov系统模型,构造含该模型的Lyapunov泛函,并利用迭代法推导满足干扰衰减度的鲁棒随机稳定判据。在此基础上,利用Schur补引理,获取满足H?范数界的线性矩阵不等式(linear matrix inequality,LMI),并结合系统各子工况最小方差约束,设计满足最小方差约束的鲁棒随机稳定控制器。然后,利用线性目标函数的最小化问题求解各子工况匹配控制器,利用随机梯度在线辨识方法确定子工况对应加权系数,实现自适应加权控制。时域仿真结果表明该策略具有良好的控制效果,有效解决了固定故障集无法与当前工况合理匹配的难题。An adaptive control strategy based on discrete Markov power system model with multiple operating conditions is proposed based on cascading failure causing large variation in operating condition of power system. First, the discrete Markov power system model considering multi-operation condition factor is established and Lyapunov functional containing this model is constructed. Iterative method is used to deduce robust stochastic stability criterion meeting interference decay rate. On this basis, Schur complement lemma is used to obtain linear matrix inequalities （LMI） meeting H～ norm bound and robust stochastic stabilization controller meeting minimum variance constraints combined with constraints of various operating conditions of system is designed. Then minimization problem of linear objective function is used to solve the matching controller for each operating condition and weighted coefficient of each condition is determined on line with stochastic gradient method to realize adaptive weighted control. Time-domain simulation results show that the strategy has good control effect and solves effectively the problem that fixed fault sets cannot be matched with current operating condition reasonably.

关 键 词：离散马尔可夫系统模型 多工况系统 在线辨识 鲁棒随机稳定 自适应控制策略 

分 类 号：TM71[电气工程—电力系统及自动化]