Field Oriented Control

AC Induction motors are robust and widely used, but precise torque control is harder than simple steady-state speed control. Rotor flux, slip, magnetic saturation, temperature-dependent resistance and inverter limits affect the dynamic model. Field-oriented control uses that model to regulate the motor’s current vector.

Introduction of Field Oriented Control

Scalar V/Hz control sets voltage magnitude as a function of frequency and suits applications with modest dynamic requirements. Field Oriented Control (FOC) adds current feedback, a rotor or flux angle and a motor model to control transient torque and flux more directly. Digital controllers make the required transformations and control loops practical.

FOC makes an AC drive behave like a separately excited DC motor by expressing stator current in a rotating reference frame. When the chosen flux vector defines the d-axis, one stator current component mainly controls flux and the orthogonal component mainly controls torque. This decoupling is approximate and depends on angle estimation, machine parameters, current-loop bandwidth and compensation of cross-coupling terms.

FOC is used in appliances as well as high-performance drives. It can provide smooth torque and fast acceleration over a wide speed range. Full torque at zero speed is possible only when rotor position or flux angle remains observable, current and thermal limits allow it, and the inverter is designed for sustained low-speed current.

Working Principle of Field Oriented Control

Field oriented control regulates the stator-current vector in a rotating d-q reference frame. The Clarke transform maps three phase quantities to stationary α-β coordinates, and the Park transform rotates them into d-q coordinates using an electrical angle. In balanced steady state, correctly aligned d-q values can become nearly constant, which simplifies PI current control. The q coordinate is commonly assigned to torque-producing current and the d coordinate to flux-producing current, but the exact torque relation depends on motor type and axis convention.

Three-phase voltages, currents and fluxes can be represented by space vectors. Let i_a, i_b and i_c be the instantaneous stator phase currents. The stator current space vector is then defined by the equation below.

Here, a, b and c are the axes of the three phase system. Transform scaling can be amplitude-invariant or power-invariant, so implementations must use one convention consistently.

This current space vector represents the three-phase system. FOC transforms it in two steps:
(a, b, c) → (α, β) is the Clarke transformation into a stationary orthogonal frame.
(α, β) → (d, q) is the Park transformation into a rotating orthogonal frame.
The (a, b, c) → (α, β) Projection (Clarke transformation)
The Clarke transform maps three-phase voltages or currents onto stationary α and β axes. The α-β components remain time-varying for a rotating sinusoidal vector. One common transformation matrix is shown below:

Assume that phase axis a and stationary axis α point in the same direction and that β is orthogonal to α. The vector diagram then uses the following geometry:

The Clarke projection maps the three phase system into the stationary two-dimensional α-β system shown below:

The α-β currents still vary with time as the current vector rotates.
The (α, β) → (d.q) projection (Park transformation)
The Park transform rotates the stationary α-β components into the d-q reference frame. Its matrix is shown below:

Here, θ is the electrical angle from the stationary frame to the rotating frame.
If rotor flux defines the d-axis, the diagram shows both representations of the current vector:

The angle θ must track the chosen flux vector. The d-q components of the current vector follow from the equations below:

The calculation depends on the α-β components and an accurate electrical angle. In rotor-flux-oriented induction-motor control, i_sd mainly sets rotor flux and i_sq mainly sets torque. Actual decoupling also requires machine-model terms and current regulators; the two axes are not physically independent under every transient or saturation condition.

Basic Module for Field Oriented Control

Current sensors measure two or three stator phase currents. The Clarke block produces i_sα and i_sβ. The Park block uses the electrical angle to convert them into d-q current components i_sd and i_sq. Inner current controllers compare these values with i_sdref and i_sqref. The reference generator depends on the machine and operating region. For a surface-mounted PMSM below base speed, i_sdref is often zero. In induction motors, magnetising current requires a non-zero d-axis reference. Interior-PM machines can also use a non-zero d-axis reference for maximum torque per ampere, and both PMSM types can use negative d-axis current for field weakening. The current-controller outputs are V_sdref and V_sqref. An inverse Park transform produces stationary commands V_sαref and V_sβref for space-vector PWM, which generates inverter switching commands. Park and inverse Park calculations require a consistent electrical angle. Current limits, voltage limits, dead-time compensation and anti-windup are also part of a practical drive.
Synchronous and induction machines obtain the angle differently.

1. For synchronous motor(s), the rotor field is synchronous with the commanded stator field. A position sensor or observer provides mechanical angle, which is converted to electrical angle using the pole-pair count and calibrated offset. Integrating speed alone accumulates error unless the angle is periodically corrected.
2. For an asynchronous motor, rotor-flux electrical speed differs from rotor mechanical electrical speed by slip speed. An observer or indirect model estimates flux magnitude, slip and angle from measured current, voltage, rotor speed and induction motor parameters. Accuracy is sensitive to resistance, saturation and low-speed signal quality.

Simplified Indirect FOC Block Diagram

Classification of Field Oriented Control

Rotor-flux FOC for an induction motor drive is commonly described as direct or indirect. Direct FOC obtains flux angle from a flux sensor or a flux observer, often using measured current and voltage. Indirect FOC calculates synchronous angle from rotor position or speed plus modelled slip frequency. Both are forms of field oriented control, and both depend on parameter accuracy. Zero-speed torque, sensorless operation and field weakening require separate validation; indirect FOC does not guarantee them by itself.

Advantages of Field Oriented Control

1. Fast torque response when current loops and angle estimation are properly tuned.
2. Low-speed torque control with a suitable position sensor or low-speed observer.
3. Accurate dynamic speed control with a closed speed loop.
4. Efficient use of motor current and flux over the designed operating range.
5. Four-quadrant operation when the inverter and DC power path support bidirectional torque and energy flow.
6. Short-term overload only within inverter current and motor thermal limits.