An adaptive method for tuning process noise covariance matrix in EKF-based three-phase distribution system state estimation

This paper proposes a new adaptive method for online tuning of process noise covariance matrix in the Extended Kalman Filter based three-phase distribution system state estimator. Specifically, a new form of exponential function is proposed for tuning the process noise covariance matrix, adapting it to the level of state changes. A new indicator derived from normalized innovations of the available real-time and virtual measurements is developed for tracking the level of state changes. The proposed method relies on measurements of the existing distribution networks, and therefore can be easily implemented in the advanced distribution management system. The method efficiently adapts process noise to system state variations and can deal with both quasi steadystate and unexpected sudden state changes caused by changes in the system topology, generation or demand. Comparison results with other state-of-the-art adaptive methods on the modified IEEE 37-bus and IEEE 123-bus distribution systems show that the proposed method achieves better accuracy under quasi steady-state condition while being more robust to unexpected sudden state changes.

Automated summary

This study introduces a new adaptive method that adjusts the process noise covariance matrix to adapt to the level of state changes, and utilizes normalized innovations to track the level of state changes. The proposed method achieves better accuracy and robustness in distribution systems.