Sensorless Permanent Magnet Machine Control via Vector Tracking Observer
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for controlling permanent magnetic machines without mechanical position sensors are susceptible to noise, leading to inaccuracies and delays in rotor position estimation.
Innovation Solution
A system that includes a state evaluator to determine operating speed ranges, a sensor to sense back electromotive force (EMF) currents, and a vector tracking observer to estimate rotor position and motion data using heterodyning and inertial models, eliminating the need for mechanical sensors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If arc-tangent function is used to estimate rotor position from filtered position, then position estimation is achieved, but noise susceptibility increases leading to inaccuracies and delays
Solution Approach 1:
The estimation system is divided into multiple independent observers (primary Luenberger observer for position, secondary observer for velocity/acceleration, vector tracking observer for back-EMF) that process different aspects of the system separately. Each observer handles specific calculations independently, allowing optimized noise filtering for each function without compromising overall accuracy.
Solution Approach 2:
A secondary observer acts as an intermediary between the primary position estimation and the final control system. It processes the position data through additional filtering and differentiation stages, mediating the transfer of information while reducing noise propagation to the velocity and acceleration estimates.
2Measurement precision
If mechanical position sensors (encoder or resolver) are used, then accurate position data is obtained, but system complexity and cost increase
Solution Approach 1:
The system uses its own operational parameters (current, voltage, frequency) to generate position estimates without external sensing equipment. The motor's back-EMF and current characteristics serve as the sensing mechanism, eliminating the need for separate mechanical position sensors.
Solution Approach 2:
Mechanical position sensors (encoders, resolvers) are replaced with an electronic estimation system based on electrical measurements. The physical mechanical sensing components are substituted with computational algorithms that derive position from electrical signals.
3Adaptability or versatility
If back-EMF vectors are directly used for position estimation, then sensorless control is achieved, but noise susceptibility increases
Solution Approach 1:
The back-EMF vectors are pre-processed through a vector tracking observer that performs preliminary estimation and filtering before being used by the secondary observer. This preliminary action cleans the signal and prepares it for more accurate position and velocity calculation.
Solution Approach 2:
The system implements feedback loops where the estimated position and velocity are continuously refined based on the difference between predicted and actual electrical measurements. The observers use feedback from current and voltage measurements to correct and improve position estimation accuracy.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enhances the immunity to noise, providing more accurate and reliable rotor position and motion estimation, enabling effective control of permanent magnetic machines across varying speed ranges.
Implementation Method 1
A sensor is configured to sense current, of one or more output phases of an inverter, associated with back electromotive force (back EMF) of the machine
Data Source
Figure 1
Figure 2
Figure 3A~3B
AI summary
A sensor is configured to sense current, of one or more output phases of an inverter, associated with back electromotive force (back EMF) of the machine. A converter or electronic data processor is adapted to convert the sensed current into current vectors associated with a stationary reference frame. An estimator or current model is configured to estimate back-EMF vectors from the converted current vectors. A vector tracking observer or the electronic data processor is adapted to mix the back-EMF vectors and applying the mixed back-EMF vectors to a preliminary inertial model. A secondary observer or the data processor is operable to apply the output of the preliminary inertial model to a secondary inertial model in the second speed range to estimate position or motion data for the rotor.