Machining State Analysis Using Measurement-Linked Machine Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional machining analysis techniques fail to identify the specific factors causing defective machining, such as human, tool, jig, workpiece, or machine-related issues, by not effectively associating measured data from machined parts with machine data during operation.
Innovation Solution
A device and method that collect and analyze machine data and measured data from machining processes, extracting feature data to determine the cause of defective machining by associating size measurements of machined parts with machine output data, allowing for real-time identification of factors contributing to defects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional machining analysis techniques are used to detect abnormality by comparing load torque patterns, then abnormality detection capability is improved, but the ability to identify specific defect factors (human, tool, jig, workpiece, machine) deteriorates
Solution Approach 1:
The patent segments the machining data into multiple categories corresponding to different defect factors (human factor, tool factor, jig factor, workpiece factor, machine factor). By dividing the analysis into these specific segments, the system can identify which particular factor is causing the abnormality rather than just detecting that an abnormality exists. This is achieved through separate feature extraction units for each factor type and corresponding determination units.
Solution Approach 2:
The patent adds a new dimension to the analysis by incorporating measured data from measuring machines into the existing machine data analysis. This creates a multi-dimensional analysis framework where machine data (load torque, current, speed) is combined with measurement data (dimensions, shape, surface quality) to comprehensively identify defect factors across different dimensions of machining quality.
2Ease of manufacture
If machine data and measured data are not associated with each other, then data collection simplicity is improved, but the precision of machining state analysis deteriorates
Solution Approach 1:
The patent introduces a data association mechanism that acts as an intermediary between machine data and measured data. The determination units serve as intermediaries that correlate machine data (from machine tools) with measured data (from measuring machines) by matching measurement points with corresponding machine operation data. This intermediary layer enables precise association without requiring direct integration between different data sources, maintaining data collection simplicity while achieving high analysis precision.
3Device complexity
If feature data is not extracted from machine data, then data processing complexity is improved, but the accuracy of defect factor identification deteriorates
Solution Approach 1:
The patent extracts specific feature data from the raw machine data through dedicated feature extraction units. Each extraction unit focuses on extracting relevant features for a specific defect factor (e.g., extracting features related to tool wear from current and torque data). This extraction process separates the essential diagnostic information from the raw data, improving identification accuracy while managing processing complexity through targeted extraction rather than comprehensive analysis of all raw data.
Data Source
AI summary
To provide an analysis device, an analysis method, and an analysis program capable of analyzing a machining state while associating machine data output during operation of a machine tool and measured data containing the size of an actual machined part measured by a measuring machine with each other. An analysis device comprises: a collection unit that collects an aggregate of machine data output during operation of a machine tool and an aggregate of measured data containing measurement points where the size of a machined part machined by the machine tool has been measured by a measuring instrument; and a feature extraction unit that selects machine data corresponding to an arbitrary measurement point, in the aggregate of the measured data from the aggregate of the machine data, and extracts the selected machine data as a feature at the arbitrary measurement point.