What is state estimation in power systems and what data are typically used?

What is state estimation in power systems and what data are typically used?

• A. State estimation uses a simple average of all measurements to estimate voltages.
• B. State estimation uses a Weighted Least Squares optimization to estimate bus voltages and angles from heterogeneous measurements, including SCADA and PMUs, and to detect bad data.
• C. State estimation only uses SCADA data and ignores PMUs.
• D. State estimation predicts future loads without any measurements.

Answer: (B)

State estimation is the process of determining the voltages and their phase angles at all buses in the power system from measurements, rather than by direct measurement of every state. The standard method uses a Weighted Least Squares optimization to find the state vector (bus voltages and angles) that best fits all available measurements, and it performs bad-data detection through the residuals of the fit. Measurements come from two main sources: SCADA, which provides broad coverage data like power flows, injections, and voltage magnitudes; and PMUs, which deliver time-synchronized phasors—both magnitude and angle—for voltages and currents at high speed. Because these measurements have different accuracies and update rates, the estimator weights them accordingly, giving more influence to the more reliable data. The relationships between measurements and the state are nonlinear, so the problem is solved iteratively (often with Gauss-Newton) to converge on the most plausible state. The result is an estimated set of bus voltages and angles plus residuals used to flag bad data and ensure observability. Averaging measurements would miss the physics and constraints, using only SCADA neglects valuable angle information from PMUs, and predicting loads without measurements isn’t estimating the current state at all.

State estimation is the process of determining the voltages and their phase angles at all buses in the power system from measurements, rather than by direct measurement of every state. The standard method uses a Weighted Least Squares optimization to find the state vector (bus voltages and angles) that best fits all available measurements, and it performs bad-data detection through the residuals of the fit. Measurements come from two main sources: SCADA, which provides broad coverage data like power flows, injections, and voltage magnitudes; and PMUs, which deliver time-synchronized phasors—both magnitude and angle—for voltages and currents at high speed. Because these measurements have different accuracies and update rates, the estimator weights them accordingly, giving more influence to the more reliable data. The relationships between measurements and the state are nonlinear, so the problem is solved iteratively (often with Gauss-Newton) to converge on the most plausible state. The result is an estimated set of bus voltages and angles plus residuals used to flag bad data and ensure observability. Averaging measurements would miss the physics and constraints, using only SCADA neglects valuable angle information from PMUs, and predicting loads without measurements isn’t estimating the current state at all.