Let a three-phase system (voltages, currents, flux linkages) be represented by a single complex time-varying vector in a stationary two-dimensional plane (the $\alpha\beta$-plane). For a set of phase quantities $x_a, x_b, x_c$ satisfying $x_a + x_b + x_c = 0$, the space vector is defined as:
The three-phase machine is one entity. Its state is a rotating complex number. Unbalance, harmonics, and switching states (inverters) become geometric loci, not case-by-case trigonometric expansions.
$$\frac{d\vec{\psi}_s}{dt} = \vec{v}_s - R_s \vec{i}_s$$ Let a three-phase system (voltages, currents, flux linkages)
where $\omega_k$ is the speed of the chosen reference frame (stationary, rotor, synchronous). The torque expression unifies as:
Difference between machine types is merely a matter of flux generation: $\vec{\psi}_s = L_s \vec{i}_s$ (IM), $\vec{\psi}_s = L_s \vec{i} s + \vec{\psi} {PM}$ (PMSM), or $\vec{\psi}_s = L_s \vec{i}_s + L_m \vec{i}_r'$ (DFIM). The drive —the control algorithm—does not need to know the difference beyond the flux linkage map. The drive —the control algorithm—does not need to
“The space vector is not a mathematical trick. It is the machine’s own memory of what it is.”
The art of modern drive control (field-oriented control, direct torque control, model predictive control) reduces to selecting, in real time, the inverter switching state that minimizes a cost function of the flux and torque errors. No sinewave mythology required. flux drift—is simply the integral of:
When coupled to a voltage-source inverter, the space vector approach reveals the finite set of discrete stator voltage vectors ($V_0$ to $V_7$). The machine’s response—current trajectory, torque ripple, flux drift—is simply the integral of: