Machine Learning Chatter Prediction for Machining Condition Adjustment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for machining with machine tools often result in chatter vibration, which affects machined surfaces, and require manual adjustment of spindle rotational frequency and feed speed to prevent, but this process is inefficient and relies on trial and error.
Innovation Solution
A chatter vibration determination device using machine learning to analyze machining condition data, including feed speed and spindle rotational frequency, to predict and prevent chatter vibration by adjusting machining conditions before it occurs, utilizing learning models to estimate vibration occurrence and provide adjustment guidance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual adjustment of machining conditions is performed to prevent chatter vibration, then chatter vibration can be reduced, but the process is inefficient and relies on trial and error
Solution Approach 1:
The learning model predicts chatter vibration occurrence in advance based on machining condition data before actual machining occurs. This allows workers to adjust machining conditions proactively to prevent chatter vibration, rather than reacting after it occurs, thereby improving both reliability and productivity
Solution Approach 2:
The system collects machining condition data and uses a learning model to provide feedback predictions about chatter vibration occurrence. This feedback loop enables continuous optimization of machining conditions based on predicted outcomes, eliminating trial-and-error approaches
2Reliability
If machining conditions are adjusted after chatter vibration occurs, then chatter vibration can be reduced, but substantial chatter vibration may already have affected machined surfaces
Solution Approach 1:
The learning model performs prediction before chatter vibration actually occurs during machining. By analyzing machining condition data in advance, the system alerts workers to adjust conditions before substantial vibration affects the workpiece, preserving surface quality while maintaining vibration control
3Reliability
If trial and error method is used to determine optimal machining conditions, then chatter vibration can be minimized, but time is wasted and productivity decreases
Solution Approach 1:
The patent replaces the mechanical trial-and-error adjustment process with an information-based machine learning prediction system. The learning model analyzes machining condition data and provides predictive guidance, substituting empirical trial-and-error with computational prediction, thereby eliminating time loss while achieving vibration minimization
Data Source
AI summary
A chatter vibration determination device is provided with a machine learning device configured to observe machining condition data including a feed speed and a spindle rotational frequency in cutting as state data representative of the current state of environment, execute processing related to machine learning using a learning model obtained by modeling the relationship of chatter vibration with a machining condition for the cutting, based on the state data, and estimate the occurrence/non-occurrence of chatter vibration and the improvement of the chatter vibration. The chatter vibration determination device outputs the result of the estimation of the occurrence/non-occurrence of the chatter vibration and the improvement of the chatter vibration.