A mixed integer robust state estimator for smart distribution network

This paper propose a robust state estimation method for smart distribution network Integer variables are introduced to detect and rule out bad data automatically. Measurements from root bus are given the most weight by incorporate them as constant in power flow constraints. The proposed method is a mixed integer nonlinear problem (MINLP). Complementary theory is adopted to transform the MINLP into NLP, which balances solving speed and accuracy. A modified version of a public network is used as test benchmark. Three cases with DGs are simulated and discussed deeply. Simulation shows that the proposed method can be implemented on both over-estimated and under-estimated network Error of estimated value is still within acceptable range even in the worst case in which there are hardly measurements. The proposed method is fast when solved with interior point method and is capable of on-line state estimation.