Machine Learning Device for Servo Motor Compensation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In servo control devices with feedback loops, inappropriate compensation values can lead to vibration and machine shutdowns during machine learning, disrupting the learning process.
Innovation Solution
A machine learning device that includes an abnormality detection unit to stop compensation and continue optimization, using reinforcement learning to adjust compensation values and prevent inappropriate selections, with a switch to isolate the feedback loop and a low-pass filter for smooth transitions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If reinforcement learning is used to machine-learn coefficients of a transfer function while actually operating the machine tool, then learning effectiveness is improved, but inappropriate compensation values may cause vibration or machine shutdown interrupting learning
Solution Approach 1:
The system implements a feedback mechanism where the learning control unit monitors compensation values generated during reinforcement learning and compares them against predetermined appropriate ranges. When compensation values fall outside this range, the system detects this as an abnormal state and interrupts the learning process, preventing vibration and machine shutdown. This feedback loop ensures learning continues reliably while maintaining operational stability.
2Reliability
If compensation data is generated by repeating idle operation, then vibration and shutdown are prevented, but learning efficiency is reduced compared to actual operation
Solution Approach 1:
The system dynamically adapts the learning mode based on compensation value validity. It transitions between two operational states: (1) actual machine tool operation for efficient learning when compensation values are within appropriate ranges, and (2) idle operation when compensation values exceed appropriate ranges. This dynamic switching optimizes learning efficiency while maintaining reliability, combining the advantages of both idle and actual operation methods.
3Stability of the object's composition
If a switch is added to isolate the feedback loop and a low-pass filter for smooth transitions, then system stability is improved, but device complexity increases
Solution Approach 1:
The system introduces a switch as an intermediary component between the learning control unit and the feedback loop. This switch acts as a mediator that can isolate the feedback loop when compensation values are inappropriate, preventing instability and vibration. The low-pass filter serves as another intermediary to smooth transitions during mode switching. These intermediary elements add minimal complexity while significantly improving system stability and learning continuity.
Data Source
AI summary
Provided is a machine learning device configured to perform machine learning related to optimization of a compensation value of a compensation generation unit with respect to a servo control device configured to control a servo motor configured to drive an axis of a machine tool, a robot, or an industrial machine, and that includes at least one feedback loop, a compensation generation unit configured to generate a compensation value to be applied to the feedback loop, and an abnormality detection unit configured to detect an abnormal operation of the servo motor, wherein, during a machine learning operation, when the abnormality detection unit detects an abnormality, the compensation from the compensation generation unit is stopped and the machine learning device continues optimization of the compensation value generated by the compensation generation unit.