Magnetic Bearing Control Learning for Stable Shaft Levitation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Magnetic bearing devices face challenges in maintaining precise control of the position of supported objects due to product-to-product variation and temporal changes, leading to instability in levitation and potential contact with touchdown bearings.
Innovation Solution
A machine learning device is integrated into the magnetic bearing system, utilizing a learning unit that acquires state variables and evaluation data to adjust control conditions for the magnetic bearing, including voltage and current values of electromagnets, to optimize position control and reduce power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional control methods are used for magnetic bearing devices, then the basic levitation function is maintained, but position control stability deteriorates due to product-to-product variation and temporal changes
Solution Approach 1:
The magnetic bearing device performs self-learning by automatically acquiring state variables during operation, evaluating control accuracy, and updating its own control parameters through the learning unit. This enables the system to adapt to product-to-product variations and temporal changes without external intervention, thereby maintaining reliable position control stability.
Solution Approach 2:
The system implements continuous feedback by acquiring state variables (position, current, voltage) and evaluation data (control accuracy metrics) during operation. The learning unit uses this feedback information to iteratively update control parameters, enabling the system to compensate for variations and maintain stable position control over time.
2Reliability
If machine learning is applied to adapt to variations, then position control stability is improved, but device complexity increases due to additional learning units and data processing
Solution Approach 1:
The control unit is designed to perform multiple functions: it serves as both the conventional controller for basic levitation and the learning unit for adaptive optimization. By integrating these functions into a single control unit, the system achieves improved levitation stability through machine learning without significantly increasing overall device complexity.
Solution Approach 2:
The patent combines the learning unit, state variable acquisition unit, evaluation data acquisition unit, and updating unit into an integrated control system. This merging of functions reduces the number of separate components and simplifies the overall device structure while still achieving adaptive position control stability.
3Measurement precision
If learning data is continuously acquired and processed, then control accuracy is improved, but loss of time occurs due to data processing requirements
Solution Approach 1:
The system performs preliminary data processing by continuously acquiring and storing state variables and evaluation data during normal operation. This preliminary accumulation of data allows the learning unit to perform batch processing and parameter updates at optimal intervals, rather than requiring real-time processing of every data point, thereby reducing time loss while maintaining high measurement precision.
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 machine learning device enhances the stability of position control, reducing the likelihood of contact with touchdown bearings and improving levitation stability by adapting to variations and changes in the system.
Implementation Method 1
a magnetic bearing having a plurality of electromagnets that apply an electromagnetic force to a shaft
Data Source
AI summary
A machine learning device learns a control condition for a magnetic bearing device that includes a magnetic bearing having a plurality of electromagnets that apply an electromagnetic force to a shaft. The machine learning device includes a learning unit, a state variable acquisition unit, an evaluation data acquisition unit, and an updating unit. The state variable acquisition unit acquires a state variable including at least one parameter correlating with a position of the shaft. The evaluation data acquisition unit acquires evaluation data including at least one parameter selected from a measured value of the position of the shaft, a target value of the position of the shaft, and a parameter correlating with a deviation from the target value. The updating unit updates a learning state of the learning unit by using the evaluation data. The learning unit learns the control condition in accordance with an output of the updating unit.