Subsynchronous Oscillations (SSO) Technical Primer

Start here

What is a subsynchronous oscillation?

Every power system runs at a fixed frequency — 50 Hz or 60 Hz. A subsynchronous oscillation (SSO) is any electrical or mechanical oscillation that sustains itself below that frequency. The key point is that SSO is defined by where on the frequency scale the oscillation sits, not by what causes it.

SSO — working definition

Oscillations with a frequency below 50 Hz or 60 Hz (the system's synchronous frequency). Multiple physically different phenomena carry this label — they share a frequency range but have completely different causes.

SSO specifically excludes one type of oscillation you may already know — the slow rotor-angle swings of synchronous machines (inter-area and local modes in the 0.1–2 Hz band). Those are a separate class of problem governed by different physics and are studied independently. SSO deals with the faster band sitting between those slow swings and the 50 Hz or 60 Hz fundamental.

Why has SSO become more important? For a long time it was mainly a concern for series-compensated transmission lines and a handful of HVDC links. Modern grids have changed that: wind farms, solar plants, HVDC converters, and FACTS devices all contain fast control loops that can interact with the network — and with each other — creating IBR-driven oscillation events that older analysis methods were never designed to catch.

The mental model

Two families, one frequency band

The clearest way to understand SSO is to split it into two top-level groups based on what causes it. Every specific phenomenon in the classification falls into one of these two groups.

Family 1

Subsynchronous Resonance (SSR)

A resonance in the electrical network drives the oscillation — usually created by a series capacitor. The oscillation may stay purely electrical, or it may couple into the mechanical shaft of a turbine-generator.

Family 2

Power-Electronic Device Interactions (PEDI)

Fast control loops inside power-electronic converters — acting alone or reacting to each other — create an unstable oscillation. No series capacitor or turbine shaft is involved.

Cross-cutting

Electrical vs. Torsional

For SSR, the most important question is whether the oscillation stays in the electrical network or also drives the turbine-generator shaft into physical vibration — because shaft vibration can cause metal fatigue, which is the most severe outcome in SSO.

The full classification

How the phenomena fit together

The diagram below maps SSO into its two families, separates resonance into electrical and torsional forms, and names each specific phenomenon. Use it as a reference while reading the detailed sections that follow.

Figure 1 — Classification of subsynchronous oscillations.

Family 1 · Resonance

Subsynchronous resonance, explained

SSR occurs when an electrical resonance in the network falls below the system frequency. The classic cause is a series capacitor used to compensate long transmission lines — the capacitor and line inductance together form an LC circuit that resonates below 50 Hz. SSR appears in two forms.

Electrical SSR

Only electrical quantities — voltages and currents — oscillate. The turbine-generator shaft does not move. The energy stays within the network and machine windings.

Induction Generator Effect

Self-excitation at subsynchronous resonance

IGE

At subsynchronous frequencies, a rotating machine looks like a negative resistance to the network. If that negative value is larger in magnitude than the circuit's positive resistance at the resonant frequency, the oscillation grows rather than dying away. This mechanism was identified by Concordia and Carter in 1941 and is sometimes called self-excitation.

Network Resonance

Including Wind Subsynchronous Control Interaction (Wind-SSCI)

Wind-SSCI

A converter's fast current controller can add negative resistance at a network resonance frequency, making the oscillation grow instead of damping out. The 2009 ERCOT event — where a Type-3 wind farm (DFIG) was connected radially to a 75%-compensated 345 kV line — is the most-cited real-world example. The industry often calls this "subsynchronous controller interaction (SSCI)," but the oscillation lives in the network, not inside the controller, so it is correctly classified as electrical SSR whose behaviour is sensitive to controller settings.

Field case: An approximately 20 Hz current and voltage oscillation developed after a line outage created a direct DFIG-to-series-capacitor path; the oscillation persisted until the capacitor was bypassed.

Torsional SSR

Here the electrical resonance transfers energy into the mechanical shaft of the turbine-generator. From the rotor's perspective, the network resonance appears at a complement frequency: f_r = f₀ − f_er. If that complement frequency lines up with one of the shaft's natural torsional modes and the net damping goes negative, the shaft starts to vibrate — and that vibration can crack the shaft over time.

Transient Shaft Torque

Torque amplification during large disturbances

TST / TA

This is a large-disturbance (fault-triggered) mechanism. When a network fault occurs and a shaft torsional mode is close to the complement frequency, the shaft torque can swing violently in a single transient — it does not build up gradually over many cycles like the steady-state cases. This mechanism was formally documented in the 1977 Navajo Project report.

Torsional Interaction against Network

TI-N — network-driven torsional interaction

TI-N

This is the classic case. An electrical resonance in the network, when viewed from the rotor frame, falls at the same frequency as one of the shaft's torsional modes. The two reinforce each other and the oscillation grows. The 1970 Mohave incident — which cracked a turbine-generator shaft — is the defining real-world example, and the direct reason why torsional protection relays and blocking filters exist today.

Field case: At Newmont TS Power Plant, planned series compensation prompted the installation of torsional protection and field measurement of shaft mode frequencies and damping margins prior to energisation.

Torsional Interaction against Devices

TI-D — device-driven torsional interaction

TI-D

Torsional SSR without any series capacitor — the shaft vibration is driven instead by the control loop of a nearby power-electronic device. The first documented case was at Square Butte in 1980, where an HVDC converter excited a 11.5 Hz torsional mode of an adjacent turbine-generator. It has since been observed with SVCs, TCSCs, and large variable-speed drives in LNG plants and offshore platforms.

Field case: The 2015 Xinjiang event — a Type-4 wind plant produced a ≈29.6 Hz oscillation that propagated through the grid and actuated torsional shaft protection relays on nearby thermal generating units.

Family 2 · Converter-driven

Power-electronic device interactions

As HVDC converters, FACTS devices, and inverter-based generation (wind, solar) become more common, a new type of SSO has emerged that needs neither a series capacitor nor a rotating shaft. These are PEDI events — also called subsynchronous controller interactions (SSCI). They take two forms depending on what the converter control loop is interacting with.

Control interactions

Control Interaction between Device and Network

CI-N — device-to-network control interaction

CI-N

A single converter's control loop interacts badly with the AC network — most likely when the grid is weak (low short-circuit ratio, SCR). It can look similar to IGE on measurements, but the key difference is where the negative damping originates: in CI-N it comes from the device controls, not from a machine acting like an induction generator.

Field case: Hebei Province, China (2012–13) — 58 SSR events at 6–9 Hz on series-compensated 500 kV lines collecting DFIG wind output, self-excited as wind speed entered an unfavourable operating range.

Control Interaction between Multiple Devices

CI-D — device-to-device control interaction

CI-D

Two or more converter-based devices interact with each other through the network. This tends to occur in weak-grid conditions where high control gains — for example on an SVC, STATCOM, or wind turbine controller — make oscillations poorly damped or entirely unstable.

Field case: The 2017 Jingxia event (NW China) — a ≈37.5 Hz voltage oscillation in which DFIG and STATCOM control parameters reinforced each other until protection tripped the plant.

Reference

Which components are participating in the interaction?

For any SSO phenomenon, you need to know which components are on each side of the interaction. The table below identifies them — useful when scoping a study and deciding what needs to be modelled in detail.

Table 1 — SSO classification in terms of interacting components.
Family	Phenomenon	Component A	Component B
SSR Electrical	Induction Generator Effect (IGE)	Series capacitors and other LC resonant circuits (filters, line shunt compensation)	Synchronous & asynchronous machines (generators, motors); WTG Type 1–3
	Network Resonance		Power electronics (converters and controls)
	Wind-SSCI	Series capacitors	WTG Type 3
SSR Torsional	Transient Shaft Torque / Torque Amplification	Synchronous machines (generators, motors); WTG Type 1–3 (shaft)	Series capacitors (transient)
	Torsional Interaction against Network (TI-N)		Series capacitors (steady state)
	Torsional Interaction against Devices (TI-D)	WTG Type 1–3 (shaft)	Power electronics (converters and controls)
PEDI	Control Interaction: Device → Network (CI-N)	Power electronics (converters & controls): WTG Type 3–4, IBR, HVDC, FACTS, SVC	AC grid (weak grid)
PEDI	Control Interaction: multiple Devices (CI-D)		Other power electronics (converters and controls)

Summary

Key engineering takeaways

Key Engineering Takeaways

SSO is a frequency band, not a single mechanism. It simply means "oscillating below the system frequency." Several physically different phenomena share this label. The slow rotor-swing modes (0.1–2 Hz) are explicitly excluded.
Two families cover everything. SSR is driven by a network resonance — usually from a series capacitor. PEDI is driven by converter control loops. All the specific phenomena fall into one of these two groups.
Electrical or torsional — the distinction determines severity. Within SSR, the critical question is whether the shaft is vibrating mechanically. If it is, shaft fatigue cracking is the worst-case outcome.
The source of negative damping determines the class. IGE and CI-N can look similar on measurements, but in IGE the negative resistance comes from the machine itself; in CI-N it comes from the converter controls. Getting this right matters — the wrong diagnosis leads to the wrong fix.
Modern grids need new analysis tools. Many current SSO events involve no series capacitor at all. The fast control loops in today's converter-heavy, low-inertia networks create oscillations that traditional steady-state methods were not built to detect.

Reference. Classification, definitions, and field cases are based on CIGRE Technical Brochures.

Subsynchronous Oscillations Understanding SSO in power-electronics-dominated grids