Forced Oscillation Source Location

The primary physics-based technique for locating forced oscillation sources by calculating the net energy dissipation in each network branch. Energy flows from the source outward to the rest of the grid, and the branch exhibiting net energy injection is connected to the forcing source. The method applies the transient energy function to PMU data, computing the integral of power deviations multiplied by frequency deviations over time. Branches where energy flows into the network indicate proximity to the disturbance origin.

Forced oscillations exhibit fundamentally different characteristics from natural inter-area modes, enabling algorithmic discrimination:

• Constant frequency: The oscillation frequency matches the external driving force and does not shift with system conditions
• Persistent amplitude: Unlike decaying natural modes, forced oscillations maintain steady or growing amplitude until the source is removed
• Sharp resonance peaks: When the forcing frequency aligns with a system mode, amplitude amplifies dramatically but the frequency remains locked to the source
• Spatial localization: Energy propagates from a fixed geographic point rather than appearing as a distributed mode shape

Source location algorithms analyze the temporal evolution of energy flows across multiple PMU channels to triangulate the disturbance origin. Key processing steps include:

• Bandpass filtering to isolate the oscillation frequency of interest from ambient noise
• Energy computation using the product of branch power deviations and bus frequency deviations
• Directionality assessment to determine whether each branch is sourcing or sinking oscillatory energy
• Topological aggregation to identify the bus or generator with the highest net energy injection

The method works even when the forcing frequency coincides with a natural system mode, a condition known as resonance.

When the forcing source is a malfunctioning generator, specific signatures appear in the terminal measurements:

• Sustained power oscillations at the generator's mechanical natural frequency, often linked to turbine governor hunting or valve cycling
• Phase relationship anomalies between terminal voltage and current that deviate from normal synchronous machine behavior
• Excitation system feedback where automatic voltage regulator oscillations couple with mechanical forcing
• Shaft torsional signatures detectable through sub-synchronous components in the electrical output

These signatures allow discrimination between electrical network faults and prime mover malfunctions.

Accurate source location requires spatially distributed PMU coverage with synchronized timestamps. The algorithm correlates measurements across multiple substations:

• Cross-correlation analysis identifies the time delay of oscillation arrival at different PMU locations, with the earliest arrival indicating proximity to the source
• Coherence metrics quantify the linear relationship between oscillation components at different buses, degrading with distance from the source
• Phase angle propagation maps the spatial progression of the oscillation wavefront across the transmission network
• Observability requirements mandate PMU placement at major generation buses and tie-line interconnections for reliable triangulation

Source location algorithms feed directly into Remedial Action Schemes (RAS) and operator alerts for rapid mitigation:

• Automated source tripping isolates the forcing generator or load within seconds of confirmed location
• Operator visualization displays the energy flow direction on geographic one-line diagrams for situational awareness
• Oscillation Baselining establishes normal ambient energy levels to reduce false positives during routine grid fluctuations
• Post-event forensic analysis archives synchrophasor data with source location metadata for NERC disturbance reporting and root cause investigation

A vector describing the relative amplitude and phase participation of generators and buses in a specific oscillatory mode, critical for pinpointing the source location.

• Generators with the largest mode shape magnitude are primary participants
• Phase angle differences reveal the geographic propagation pattern of the oscillation
• Forced oscillations often exhibit a distinct mode shape compared to natural inter-area modes
• Combined with DEF methods, mode shape analysis provides corroborating evidence for source triangulation

A framework of metrics ensuring PMU measurements are valid for oscillation source location algorithms. Poor data quality leads to false source identification.

• Total Vector Error (TVE) must remain below 1% during dynamic events
• Time-alignment errors across PMUs corrupt phase angle comparisons essential for DEF calculation
• GPS signal loss or PTP synchronization failures introduce catastrophic angle drift
• Continuous data quality flagging prevents corrupted frames from entering source location algorithms

The extraction of modal properties from low-amplitude random fluctuations during normal grid operation, enabling baseline characterization before a forced oscillation occurs.

• Establishes the natural damping ratios of inter-area modes under varying load conditions
• Deviations from ambient baselines indicate the onset of a forced oscillation
• Provides the pre-disturbance reference needed to isolate the forced component from natural modes
• Enables continuous monitoring without waiting for a major disturbance or ringdown event