Machining Vibration Estimation for Surface Defect Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing machining environment estimation devices struggle to accurately distinguish between spindle and table vibrations, leading to difficulties in identifying the cause of defects in machined surfaces, especially when minute amplitude vibrations are involved, and require skilled personnel to analyze surface states for quality prediction.
Innovation Solution
A machining environment estimation device that measures spindle rotational, holder unit, and table unit vibrations, performs spectral analysis, and uses machine learning with input data including machining conditions and surface roughness data to predict defects by identifying vibration sources, allowing for precise estimation of defect causes and quality prediction before machining.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If spectral analysis using FFT is performed on vibrations, then vibration frequency data can be obtained, but it becomes difficult to separately analyze the vibration caused by spindle rotation and disturbance vibrations with minute amplitude
Solution Approach 1:
The patent segments the vibration analysis into two distinct approaches: (1) time-domain analysis using RMS values to detect overall vibration levels, and (2) frequency-domain analysis using FFT to identify specific vibration frequencies. This segmentation allows the system to separately analyze spindle rotation vibrations and disturbance vibrations with minute amplitude by examining different characteristics in different domains, thereby resolving the inability to separate vibration sources in traditional single-domain analysis.
Solution Approach 2:
The patent introduces a new dimension of analysis by combining time-domain RMS values with frequency-domain FFT spectral data. Instead of relying solely on frequency analysis, the system adds the time-domain amplitude dimension to distinguish between different vibration sources. This multi-dimensional approach enables the system to identify disturbance vibrations with minute amplitude that would be indistinguishable in frequency-only analysis, as they manifest differently in the time domain.
2Measurement precision
If vibration measurement is performed to identify vibration sources, then defect causes can be identified, but it becomes difficult to distinguish between spindle vibrations and table vibrations
Solution Approach 1:
The patent introduces multiple vibration sensors positioned at different locations (spindle, tool holder, and table) as intermediaries to capture vibrations from different spatial origins. Each sensor acts as an intermediary that isolates and measures vibrations specific to its location. By comparing the data from these intermediary sensors, the system can distinguish between spindle vibrations and table vibrations, resolving the difficulty of identifying spatial vibration origins in single-point measurement systems.
Solution Approach 2:
The patent segments the vibration measurement system into multiple independent measurement points: spindle vibration sensors, tool holder vibration sensors, and table vibration sensors. Each segment measures vibrations locally, allowing the system to attribute vibrations to their specific spatial origin. This segmentation of measurement locations enables clear distinction between spindle and table vibrations, overcoming the limitation of undifferentiated vibration measurement in traditional systems.
3Measurement precision
If inspection of all machining-finished workpieces is performed to identify vibration defects, then defect causes can be identified, but productivity decreases and skilled personnel are required
Solution Approach 1:
The patent implements preliminary vibration monitoring during the machining process itself, rather than waiting for post-machining inspection. By continuously measuring vibrations in real-time and comparing them against learned patterns from the machine learning model, the system can predict potential defects before they occur on the workpiece. This preliminary detection approach eliminates the need for inspection of all finished workpieces, maintaining high defect detection accuracy while preserving productivity and reducing dependency on skilled personnel for post-processing analysis.
Solution Approach 2:
The patent replaces the mechanical inspection process (manual examination by skilled personnel) with an automated vibration-based prediction system using machine learning. The system substitutes human expertise with an automated model that analyzes vibration data and predicts surface quality outcomes. This substitution eliminates the need for skilled personnel to inspect each workpiece while maintaining or improving defect detection accuracy, thereby resolving the contradiction between inspection thoroughness and productivity.
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
Enables precise prediction of defect occurrence on machined surfaces by distinguishing between different vibration sources, reducing the need for skilled personnel and ensuring consistent machining quality across various workpieces and machining types.
Implementation Method 1
a holder unit vibration sensor that measures a holder unit vibration, a spindle unit vibration sensor that measures a spindle unit vibration, and a table unit vibration sensor that measures a table unit vibration
Implementation Method 2
spectral analysis (FFT analysis) is performed on the three measured types of the vibration data
Data Source
AI summary
A machining environment estimation device includes a data acquisition unit that acquires vibration time-series data which indicates a machining environment of the machine tool, machining conditions in carrying out machining for a workpiece in the machine tool, measurement data of a machined surface of a machining-finished workpiece, and machined surface evaluation data, a pre-processing unit that creates vibration data and machining condition data which serve as state data, and machined surface measurement data and machined surface evaluation data which serve as label data, and a learning unit that generates a learning model which learned (a) the machined surface measurement data and (b) a machined surface evaluation result of the machining-finished workpiece, with respect to (i) a vibration state and (ii) the machining conditions in the machining environment, based on the state data and the label data.