Steady-State Open-Loop High-Frequency Micro-Oscillation Model

Problem Setup & Physical Constraints

A 2-phase (x, y), single pole-pair, non-salient permanent magnet synchronous motor is driven open-loop by high-frequency balanced sinusoidal voltages () under constant PWM/voltage settings. The rotor is unconstrained mechanically (zero friction, zero external load) but possesses a known inertial mass (J) and stable torque/back-EMF constant ().

When the probing angular frequency () is higher than the motor's mechanical break frequency, the rotor fails to entrain (synchronise) with the traveling wave. Instead, it enters a dynamic equilibrium, settling at a stable mean bias position () while executing low-amplitude, high-frequency mechanical micro-oscillations driven by the alternating quadrature current.

The control loop samples the stator currents perfectly at the average current point every cycle with an arbitrary discrete PWM/sampling period (). The discrete current updates are captured as and .


Discrete State-Space Volt-Ampere Balance Equations

1. X-Axis Grouped Equation

2. Y-Axis Grouped Equation


Unified Parameter and Scaler Components

The Mechanical Reaction Current Scale ()

The Average Electrical Current Scale ()

The Average Discrete Current Step Scale ()


Parameter Identification & The "Resistance Spike" Mechanism

When parameter identification loops process these signals under the assumption of a rigidly locked rotor (), the mechanical back-EMF contributions from the micro-vibrations do not drop to zero. Instead, they project directly onto the orthogonal and base vectors.

Because scales inversely with the probing frequency squared (), lower frequencies pass significant mechanical energy back into the electrical system via back-EMF. Parameter estimators interpret this incoming mechanical energy as excess real power dissipation, resulting in a pronounced, virtual resistance spike that rapidly attenuates to zero as scales past the mechanical tracking threshold of the rotor's inertia.

The Open-Loop Parameter Separation Trick

By treating a vibrating, unconstrained rotor as a locked system during low-frequency open-loop probing, an estimator calculates four corrupted parameters: , , , and .

Because the mechanical back-EMF distorts these values through a single spatial vector (), this unique asymmetrical footprint contains enough information to simultaneously decode the true electrical parameters (), the static rotor position (), and the electromechanical torque capability () without an encoder or a datasheet.


1. The Corrupted Measurement Framework

Your independent dual-axis parameter estimator calculates four values that blend true copper attributes with the dimensionless mechanical reaction scaling factor :


2. Step-by-Step Extraction Sequence

Step A: Isolate the True Electrical Constraint

Multiplying the inductance values by the known probing frequency shifts them to the impedance domain (ohms). Because of the cross-coupling layout of the back-EMF, subtracting the cross-axis parameter pairs completely dissolves the mechanical distortion vector:

This provides a direct, uncorrupted linear tracking relationship between your true copper winding resistance and inductance.

Step B: Deduce the Unknown Rotor Angle ()

By separating the parameter differences between the two phases, the unknown mechanical reaction amplitude cancels out entirely. Evaluating the ratio of these spatial cross-axis differences isolates the static rotor alignment angle:

Step C: Resolve True and

With the spatial angle computed, the parameter system collapses back into a basic set of linear equations. Substituting reveals the uncorrupted winding resistance:

Once is resolved, the true winding inductance is calculated instantly from the constraint established in Step A:


3. Deducing the Electromechanical Group ()

Now that the true uncorrupted electrical parameters are isolated, you can calculate the raw mechanical reaction scale directly from your original data stream:

Expanding back into its continuous structural components extracts the combined electromechanical tracking constant of your motor's torque-to-inertia ratio:

Vertically Aligned High-Frequency State Equations

X-Axis Main State Equation

Y-Axis Main State Equation

Least-Squares Cost Function