Neural Network Vector Control for PMSM Reliability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional control strategies for AC electric machines, such as PID control technologies, are inefficient and unreliable, especially under unstable and uncertain system conditions, necessitating improved control systems for permanent magnet synchronous machines (PMSMs) in applications like electric drive vehicles and wind turbines.
Innovation Solution
The implementation of artificial neural networks for vector control of AC electric machines, specifically using current-loop and speed-loop neural networks to optimize dq-control voltage and drive torque signals, respectively, with the aid of dynamic programming and backpropagation through time algorithms, to enhance performance and reliability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional PID control technology is used for vector control of AC electric machines, then the control system is simple to implement, but the control performance is inefficient and unreliable under unstable and uncertain system conditions
Solution Approach 1:
The patent replaces conventional PID control algorithms with artificial neural network-based control systems. The neural networks are trained offline to learn optimal control strategies, substituting the traditional feedback-based PID approach with a data-driven intelligent control system that adapts to uncertain conditions while maintaining reliability
Solution Approach 2:
The neural network control system is trained in advance using offline training data that represents various operating conditions. This preliminary training allows the controller to have pre-learned responses to different system states, improving reliability without requiring complex real-time calculations during operation
2Speed
If neural network vector control is implemented to improve control performance under variable conditions, then control reliability and response time are improved, but the device complexity increases
Solution Approach 1:
The neural networks are trained offline before deployment, performing the complex learning process in advance. During actual operation, the trained networks provide fast predictions and control signals without requiring real-time training computations, thus achieving fast response times while managing system complexity
Solution Approach 2:
The neural network controller is designed to be self-adapting and self-regulating based on its trained knowledge. It automatically adjusts control parameters according to the current system state without requiring manual intervention or complex external control logic, improving response time while keeping the control architecture manageable
3Reliability
If neural network vector control is implemented to improve control performance under variable conditions, then control reliability is improved, but the ease of operation deteriorates
Solution Approach 1:
The neural network control system operates autonomously, automatically adjusting its control outputs based on the current system state and its pre-learned knowledge. This self-service capability improves reliability under variable conditions while maintaining ease of operation, as the system requires minimal manual tuning or intervention during operation
Data Source
AI summary
An example method for controlling an AC electrical machine can include providing a PWM converter operably connected between an electrical power source and the AC electrical machine and providing a neural network vector control system operably connected to the PWM converter. The control system can include a current-loop neural network configured to receive a plurality of inputs. The current-loop neural network can be configured to optimize the compensating dq-control voltage. The inputs can be d- and q-axis currents, d- and q-axis error signals, predicted d- and q-axis current signals, and a feedback compensating dq-control voltage. The d- and q-axis error signals can be a difference between the d- and q-axis currents and reference d- and q-axis currents, respectively. The method can further include outputting a compensating dq-control voltage from the current-loop neural network and controlling the PWM converter using the compensating dq-control voltage.