Bayesian information fusion for state estimation in AC/DC networks with unbalanced HVDC grids

This paper proposes a State Estimation (SE) method for AC/DC networks monitored through heterogeneous data sources, based on Bayesian Information Fusion (BIF). The proposed model allows to represent bipolar High-Voltage Direct Current (HVDC) grids operating in balanced or unbalanced mode. Unbalanced operation of bipolar HVDC grids is a natural consequence of single-pole outages, and may become a relevant operating condition in offshore grids, due to longe repair times. This paper casts Forecasting-Aided State Estimation (FASE) as a generic constrained optimization problem, enabling the straightforward modeling of AC/DC converter operational limits through inequality constraints and the flexible adoption of different SE objectives. Both Weighted Least Squares (WLS) and Weighted Least Absolute Value (WLAV) objectives are employed. Results demonstrate that BIF leverages sparsely available data from synchronized measurements, providing reliable estimates in networks affected by intermittent renewable generation and actions of power electronic converters.

Highlights:

• State Estimation for multiterminal AC/DC grids with multi-conductor HVDC models.
• Forecasting-Aided State Estimation as a generic constrained optimization problem.
• Bayesian Information Fusion for state estimation with heterogeneous data sources.
• Bayesian Information Fusion with WLAV to increase resilience against data inconsistencies.
• Assessment of the value of DC synchronized measurements for state estimation.