Feature Data Comparison for Low-Load Device State Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional techniques for analyzing the state of devices on railway vehicles face challenges in processing large amounts of data, leading to increased processing loads and difficulties in data comparison.
Innovation Solution
A device analysis apparatus that includes an operation data storage unit, a feature quantity data generation unit, a first computation unit, a second computation unit, and a display unit, which generates and displays feature quantity data in a way that allows for easy comparison while preventing increased processing loads.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If feature quantity data is cut out from time-series operation data for each term (e.g., monthly comparison), then data comparison capability is improved, but processing load increases enormously
Solution Approach 1:
The patent segments the time-series operation data into fixed-time interval data (e.g., 10 seconds) and assigns unique identification information to each segment. This segmentation allows the system to compare specific operational events across different time periods without processing entire monthly datasets, thereby reducing processing load while maintaining comparison capability.
Solution Approach 2:
The system performs preliminary processing by cutting out and storing feature quantity data with identification information in advance. This pre-prepared segmented data with timestamps allows for rapid comparison when needed, avoiding the need to process raw time-series data at comparison time, thus reducing processing load while enabling precise data comparison.
2Measurement precision
If feature quantity data is cut out for detailed analysis (e.g., several seconds at departure), then analysis precision is improved, but the number of samples becomes enormous making comparison difficult
Solution Approach 1:
By segmenting operation data into fixed-time intervals with unique identification information, the patent enables precise analysis of specific events (e.g., 10-second departure segments) while making comparison easy through the identification system. Users can easily compare corresponding segments across different periods without being overwhelmed by enormous sample numbers.
Solution Approach 2:
The patent introduces identification information as an intermediary that links segmented feature quantity data across different time periods. This intermediary enables easy comparison by providing a reference system, allowing users to efficiently match and compare specific operational segments without manually handling enormous datasets.
Data Source
AI summary
A device analysis apparatus includes: an operation data storage unit storing operation data indicating an operation state of a device; a feature quantity data generation unit generating feature quantity data of the device by using the operation data; a feature quantity data storage unit storing the feature quantity data; a first computation unit generating first data indicating behavior of the feature quantity data in units of a set term by using the feature quantity data stored in the feature quantity data storage unit; a second computation unit generating second data indicating behavior of latest feature quantity data by using one or more pieces of the latest feature quantity data newer than the feature quantity data used in generating the first data, among the feature quantity data stored in the feature quantity data storage unit; and a display unit displaying the first data and the second data in one graph.