EV Traction Motor Control for Predictive Torque Ripple Suppression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Electric vehicles experience undesirable torque ripples at lower speeds due to various causes, including cogging torque, airgap flux harmonics, and mechanical imbalances, which affect ride comfort and cause vehicle oscillations, and existing solutions either compromise torque density or are costly.
Innovation Solution
A method and system that utilize a machine learning unit to predict torque ripple occurrences based on rotor values, providing throttle assist at predicted times to eliminate ripples, and update a ripple identification table with new rotor values post-rotation, employing Field Oriented Control and Supervised Machine Learning algorithms to optimize torque management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If mechanical solutions (winding configuration, stator teeth geometry, rotor barrier geometry, rotor skew) are employed to reduce torque ripple, then torque ripple is reduced, but torque density is compromised
Solution Approach 1:
The patent replaces mechanical design modifications with a control-based solution. Instead of changing the physical structure of the motor (winding configuration, stator teeth, rotor barrier), the system uses a machine learning model to predict torque ripple events and applies compensatory excitation signals through the existing control system, thereby eliminating torque ripple without affecting torque density
Solution Approach 2:
The system dynamically adjusts electrical parameters (excitation signals, current commands) based on predicted torque ripple conditions. The motor controller modifies operational parameters in real-time based on machine learning predictions, allowing torque ripple reduction without permanent structural changes that would compromise torque density
2Object-affected harmful factors
If passive damping techniques (structural reinforcements, sound attenuating materials) are used to reduce torque ripple effects, then ride comfort is improved, but cost increases and the root cause is not addressed
Solution Approach 1:
The system converts the harmful torque ripple effects into a predictable pattern that can be compensated. By using machine learning to identify torque ripple conditions and applying proactive excitation signals, the system transforms a harmful mechanical phenomenon into a controllable electrical parameter, eliminating the need for costly passive damping materials and structural reinforcements
Solution Approach 2:
The system implements a feedback loop where the machine learning model continuously predicts torque ripple events based on motor operating conditions, and the controller adjusts excitation signals accordingly. This closed-loop approach addresses the root cause of torque ripple rather than merely damping its effects, providing cost-effective ride comfort improvement
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
The solution effectively reduces torque ripples, enhancing ride comfort and motor performance, addressing mechanical imbalances, and improving vehicle handling without compromising torque density or incurring excessive costs.
Implementation Method 1
The motor controller is configured to generate a rotating magnetic field for the traction motor
Implementation Method 2
The machine learning unit is configured for predicting at least one time instant when a torque ripple will occur based upon a comparison between the set of first rotor values and a set of second rotor values in a ripple identification table
Implementation Method 3
The excitation unit is configured for performing excitation at the at least one predicted time instant based upon a throttle demand and the comparison
Data Source
Figure 1A
Figure 1B
Figure 1C
AI summary
A System and Method for Reducing Torque Ripples in Electric Vehicles In an electric vehicle (10), the system includes a traction motor (250), a motor controller (200), a machine learning (SML) unit (320) and an excitation unit (340). The motor controller (200) is configured to determine a set of first rotor values (d1, q1) that includes a first magnetic flux value (id1) and a first rotor angle value (iq1). The machine learning unit (320) includes a ripple identification table (322) including a set of second rotor values for which torque ripple does not occur. The machine learning unit (320) is configured for predicting at least one time instant when a torque ripple will occur based upon a comparison between the set of first and second rotor values in the ripple identification table (322). The excitation unit (340) is configured for performing excitation at the at least one predicted time instant based upon a throttle demand and the comparison.