Motor Parameter Estimation Using Adaptive Digital Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
AC motors often operate without optimized current control loops due to unknown motor parameters like inductance, resistance, and magnetization, leading to suboptimal performance and potential fault detection issues.
Innovation Solution
A method using an adaptive filter to determine motor parameters by oversampling current data, applying a fitting function, and filtering out transient components to calculate the slope and offset of the current, allowing for precise estimation of inductance, resistance, or magnetization, thereby optimizing the current control loop and enhancing fault detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the slope of winding current is used for inductance estimation, then inductance can be determined, but the measurement is prone to noise at high frequency and electrical noise from the power inverter couples into the feedback path causing errors
Solution Approach 1:
The patent introduces an intermediary processing system that includes oversampling circuitry and digital filtering algorithms. These intermediaries process the noisy current measurements to extract accurate inductance values without being directly affected by the high-frequency noise and electrical interference from the power inverter.
Solution Approach 2:
The patent creates multiple copies of the current measurement signal through oversampling, then processes these copies through digital filtering to produce a cleaned version that accurately represents the underlying inductance characteristics without the noise contamination present in the original signal.
2Adaptability or versatility
If motor parameters are unknown, then the motor can be used in various applications, but it is not possible to optimize tuning of the current control loop
Solution Approach 1:
The patent implements a self-service mechanism where the motor control system automatically determines its own parameters (inductance, resistance, magnetization) through the described measurement and processing methods. This eliminates the need for external parameter specification while enabling optimized control loop tuning, making the system both versatile and easy to operate.
Solution Approach 2:
The patent employs feedback mechanisms where the measured current and voltage data are continuously processed to determine motor parameters, which then feed back into the control system to optimize the current control loop tuning dynamically, maintaining both versatility and ease of operation.
3Device complexity
If current is sampled once per PWM period for fault protection, then the system is simple, but the slope of current is prone to noise and not suitable for accurate inductance estimation
Solution Approach 1:
The patent applies periodic action by oversampling the current at multiple points within each PWM period rather than once per period. This periodic oversampling approach maintains a structured timing pattern while gathering sufficient data points to enable accurate inductance estimation through subsequent digital processing.
Data Source
AI summary
The present disclosure provides a system and method to determine at least one parameter of a motor, such as an inductance. The inductance can be determined, such as based on one or more digitally sampled motor winding currents. A digital filter can be applied to the digital samples such as to determine a slope of the motor winding currents. The digital filter can include a least squares fit that can be applied to the digital samples, such as to determine a slope of the motor winding currents. The least squares fit can be determined based on a computation of central tendencies such as an average value of time, an average value of current, an average value of the square of time, and the average value of the product of time and current. The average values can be determined recursively to provide improved computational speed.