Waveform Classification With Harmonic-Band Learning for Product Inspection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing non-defective product inspection methods face challenges due to individual differences in products and facilities, leading to inaccurate evaluations and decreased determination accuracy, particularly when using frequency analysis methods like FFT, which fail to account for harmonic components and noise levels.
Innovation Solution
A classification device and method that utilize a learning model to analyze time-axis waveform data, convert it into frequency characteristic data, divide it into sections, calculate maximum values, and approximate these sections to classify products based on learned data, thereby accounting for individual differences and improving accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If frequency analysis methods like FFT are used for non-defective product inspection, then the inspection process can be automated, but the determination accuracy decreases due to individual differences in products and facilities
Solution Approach 1:
The patent transforms the raw vibration waveform into frequency domain data through FFT analysis, then extracts specific frequency components (harmonic components) as features. By changing the parameter representation from time-domain to frequency-domain and selecting specific frequency bands, the system achieves accurate classification despite individual differences in products and facilities
Solution Approach 2:
The patent introduces a learning model (classification device) as an intermediary between the vibration data and the non-defective product determination. This learning model learns the relationship between frequency characteristics and product quality from training data, enabling accurate determination while accounting for individual differences without requiring manual threshold setting
2Ease of manufacture
If conventional frequency analysis is used, then the inspection method is simple to implement, but it fails to account for harmonic components and noise levels, reducing accuracy
Solution Approach 1:
The patent divides the frequency spectrum into multiple frequency bands and extracts harmonic components at different frequencies separately. By segmenting the frequency analysis into specific bands and identifying harmonic components in each band, the system captures detailed vibration characteristics that conventional methods miss, improving accuracy while maintaining implementation simplicity
Data Source
AI summary
A classification device includes: an acquisition unit that acquires time-axis waveform data of an object; a conversion unit that converts the time-axis waveform data into first frequency characteristic data; a spectrum calculator that divides the first frequency characteristic data into division sections by a predetermined bandwidth and calculates a maximum value of a spectrum for each of the division sections; an approximation processing unit that outputs second frequency characteristic data obtained by approximating the first frequency characteristic data on the basis of the maximum value of the spectrum for each of the division sections; a generator that generates third frequency characteristic data from the second frequency characteristic data by using a learning model; and a classification unit that classifies the object on the basis of the second frequency characteristic data and the third frequency characteristic data, in which the learning model is a model that has learned approximated learning data.