Waveform Pattern Learning for Early Device State Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Frequency analysis of time series data from devices like bearings and motors struggles to detect slight changes in waveform shape, making it difficult to estimate device states early from local waveform changes.
Innovation Solution
An information processing apparatus that learns a local waveform pattern and state estimator by dividing time series data into sub-time series segments based on a base cycle, using methods like One Class Learning Time-series Shapelets, to identify and detect abnormalities in device states.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If frequency analysis is performed on time series data to estimate device state, then the spectrum can be captured, but slight changes in waveform shape cannot be detected
Solution Approach 1:
The patent divides time series data into multiple sub-time series data segments based on base cycles. By segmenting the data and analyzing local waveform patterns within each segment, the method preserves local shape information that would be lost in traditional frequency analysis, enabling detection of subtle waveform changes while maintaining spectral analysis capabilities.
2Reliability
If traditional frequency analysis is used, then computational simplicity is maintained, but early detection of device abnormalities is difficult
Solution Approach 1:
The patent performs preliminary learning to extract local waveform patterns from training data before actual device monitoring. This pre-processing step creates a reference model of normal waveform characteristics, enabling early detection of abnormalities by comparing new data against this established pattern, thus improving reliability without excessive operational complexity.
Solution Approach 2:
The patent introduces local waveform patterns as an intermediary representation between raw time series data and device state estimation. These patterns serve as a mediator that captures essential waveform characteristics, enabling more sensitive detection of device abnormalities while managing computational complexity through pattern-based analysis.
Data Source
AI summary
According to one embodiment, an information processing apparatus includes a processor. The processor is configured to learn a local waveform pattern and a state estimator used to estimate a state of a device, based on multiple elements of first sub-time series data divided from first time series data representing the waveform based on a base cycle of the waveform of a physical quantity changing in accordance with an operation of a device.