Coupled Dynamic–Static Decentralized Robust State Estimation for Integrated Electricity-Gas Systems

A coupled dynamic-static decentralized robust state estimation (RSE) methodology is proposed for integrated electricity-gas systems (IEGSs). Firstly, a discrete time state space model is established for natural gas systems considering pipeline dynamics and spatiotemporal boundary conditions. Then a coupled dynamic-static RSE framework is constructed, which performs local Huber M-based static RSE for power systems with fast dynamics, and local Huber M Kalman filter (KF)-based dynamic RSE for natural gas systems with slow dynamics. A constrained iteratively reweighted least squares algorithm is proposed to uniformly solve the equality constrained RSE model and the batch-form robust linear regression model in the robust KF. Based on local prior estimates, estimation error covariance matrices and coupling constraints, an upper-level SE model is constructed to eliminate boundary mismatch and obtain posterior globally consistent estimates. The upper-level SE is converted into a non-convex distributed formulation through network partitioning, which is then solved using an improved ALADIN algorithm in a fully decentralized manner. A multi-time-scale SE scheme is also proposed to handle the measurement deficiency issue caused by different measurement refreshing rates. Comparison with existing decentralized SE methods on a typical IEGS verifies the effectiveness of the proposed methodology on estimation accuracy, robustness and convergence performance.