Phasor Estimation Algorithm

Defines the algorithm's ability to track a phasor during power system swings. Governed by IEEE C37.118, this characteristic measures latency and overshoot.

• P-Class (Protection): Prioritizes raw speed with minimal filtering, typically reporting in < 2 cycles. Used for fast relay decisions.
• M-Class (Measurement): Prioritizes accuracy and alias rejection over speed, using longer observation windows. Essential for Wide-Area Monitoring Systems (WAMS).
• Transient Response: The algorithm must avoid ringing or overshoot when the waveform undergoes a sudden magnitude or phase step change.

The ability to accurately estimate the phasor when the system frequency deviates from the nominal 50/60 Hz. Standard Discrete Fourier Transforms (DFTs) suffer from spectral leakage during off-nominal conditions, causing oscillations in the estimated magnitude and angle.

• Leakage Errors: Occur because the sampling window is no longer an integer multiple of the signal period.
• Mitigation Techniques: Advanced algorithms use re-sampling, windowing functions (like Hanning), or frequency tracking loops to adjust the sampling interval dynamically.
• Impact on TVE: Poor off-nominal rejection directly degrades the Total Vector Error, making the measurement unreliable for frequency control.

The algorithm's capacity to reject non-fundamental frequency components introduced by non-linear loads (like inverters and arc furnaces).

• Harmonic Distortion: Integer multiples of the fundamental frequency (e.g., 3rd, 5th harmonic).
• Inter-Harmonics: Frequencies that are non-integer multiples of the fundamental, common in doubly-fed induction generators.
• Filter Bank Design: A standard DFT naturally rejects integer harmonics, but advanced FIR band-pass filters are required to suppress inter-harmonics and prevent aliasing that corrupts the fundamental phasor estimate.

The ability to accurately extract the AC phasor in the presence of an exponentially decaying DC component. This offset appears in current waveforms immediately following a fault due to the inductive nature of the grid.

• Error Source: The decaying DC has a broad frequency spectrum that leaks into the fundamental frequency bin of the DFT, causing a spurious oscillation in the estimated phasor.
• Mimic Filtering: A common analog or digital high-pass filter applied to current inputs to remove the DC offset before the estimation algorithm.
• Algorithmic Correction: Advanced methods estimate the DC time constant and magnitude in real-time and subtract it from the sampled data stream to prevent saturation of the measurement.

The processing power required to execute the algorithm in real-time on embedded hardware. PMU processors have finite clock cycles and memory.

• Recursive vs. Non-Recursive: Recursive DFT implementations are computationally efficient but can suffer from numerical instability due to accumulated round-off errors.
• Fixed-Point Arithmetic: Algorithms must be stable when implemented on fixed-point digital signal processors (DSPs) to avoid overflow and quantization noise.
• Latency Budget: The total execution time of the algorithm must fit within the strict reporting latency budget (e.g., 16.67 ms for a 60 Hz system) to ensure time-synchronized output.

The precision of the derived frequency derivative, which is highly sensitive to noise. ROCOF is critical for inertia estimation and anti-islanding detection.

• Noise Amplification: Differentiation amplifies high-frequency noise and quantization errors present in the phase angle estimate.
• Smoothing Trade-off: Heavy filtering improves ROCOF accuracy but introduces unacceptable group delay for protection applications.
• Direct Estimation: Advanced algorithms estimate frequency and ROCOF directly from the waveform samples using demodulation techniques, bypassing the error-prone differentiation of the phase angle.

The primary performance metric for evaluating a Phasor Estimation Algorithm under dynamic conditions. TVE combines both magnitude error and phase angle error into a single scalar percentage. The IEEE C37.118 standard mandates a TVE below 1% during steady-state conditions. Algorithms must maintain low TVE during frequency excursions, amplitude modulation, and power swings to be considered compliant.

The mathematical backbone of nearly all Phasor Estimation Algorithms. The DFT decomposes a sampled time-domain waveform into its constituent frequency components. A recursive DFT implementation updates the phasor estimate with each new sample, minimizing computational latency. The algorithm must compensate for spectral leakage caused by off-nominal frequency operation, typically through windowing functions or frequency tracking loops.

A derived quantity computed directly from the frequency estimates produced by the Phasor Estimation Algorithm. ROCOF measures the first derivative of system frequency and is critical for inertia estimation and fast load shedding schemes. Accurate ROCOF calculation is challenging because differentiation amplifies noise; advanced algorithms employ Kalman filtering or polynomial fitting to smooth the estimate without introducing unacceptable delay.

The defining interoperability standard that specifies the steady-state and dynamic performance requirements for Phasor Estimation Algorithms. It defines two performance classes: P-class (protection, fast response) and M-class (measurement, high accuracy). The standard prescribes compliance tests including frequency ramp, amplitude modulation, and phase modulation to ensure algorithms from different vendors produce consistent results during grid disturbances.

An optimal recursive state estimator increasingly used as an alternative to the static DFT for Phasor Estimation. Unlike the DFT, a Kalman filter explicitly models the dynamic evolution of the phasor, frequency, and ROCOF as a state vector. This allows it to track rapidly changing conditions during electromechanical oscillations with lower latency and better noise rejection, making it ideal for transient stability monitoring applications.