Robot Rotating Component Diagnosis Using Filtered Vibration Spectra
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current robot diagnosis methods are inaccurate and unconvincing as they rely solely on historical data for local robot failures and require sharing of raw data, which may contain business secrets, and do not effectively detect sub-component failures in rotating components.
Innovation Solution
A method involving preprocessing signals from rotating components to filter motion information, sending preprocessed signals or spectrum information to a server for diagnosis, and using frequency amplitude analysis to detect sub-component failures, while protecting user data by masking unnecessary information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If raw data of local robots is sent to a remote server for monitoring, then remote monitoring capability is improved, but user data privacy and security deteriorate
Solution Approach 1:
The patent extracts only the necessary diagnostic features (vibration frequency, amplitude, spectral characteristics) from the raw robot operation data, rather than transmitting the complete raw data set. This extraction process removes unnecessary information that could reveal business secrets while retaining the essential diagnostic content needed for remote monitoring and fault detection.
Solution Approach 2:
The patent segments the raw vibration signal into meaningful frequency components through spectral analysis, separating diagnostic information (frequency amplitude characteristics of sub-components) from other operational data. This segmentation allows selective transmission of only the diagnostically relevant features to the remote server.
2Device complexity
If historical data of local robot is used for diagnosis, then diagnosis process is simplified, but diagnosis accuracy deteriorates
Solution Approach 1:
The patent performs preliminary signal processing and feature extraction locally before transmission, preparing the data in advance for remote analysis. By pre-processing the vibration signals to extract frequency domain characteristics and sub-component specific features, the system simplifies the remote diagnosis process while ensuring high diagnostic accuracy through optimized data preparation.
Solution Approach 2:
The patent replaces simple historical data comparison with advanced spectral analysis and frequency domain processing. By substituting basic time-domain historical data review with sophisticated frequency amplitude analysis and sub-component identification algorithms, the system achieves higher diagnostic accuracy while maintaining a streamlined remote diagnosis workflow.
3Measurement precision
If full motion information is transmitted for diagnosis, then diagnostic accuracy is improved, but data transmission efficiency and privacy protection deteriorate
Solution Approach 1:
The patent extracts only the essential diagnostic features (frequency amplitude of sub-components, spectral characteristics) from the complete motion information, eliminating redundant data before transmission. This extraction maintains diagnostic accuracy by preserving critical fault indicators while dramatically reducing data volume for efficient transmission.
Solution Approach 2:
The patent segments the comprehensive motion information into distinct frequency components and sub-component characteristics, transmitting only the segmented diagnostic features rather than the complete motion data set. This segmentation enables targeted transmission of essential information while filtering out unnecessary details.
4Measurement precision
If comprehensive motion information is analyzed, then sub-component failure detection is improved, but computational complexity and data processing time deteriorate
Solution Approach 1:
The patent replaces comprehensive time-domain motion information analysis with frequency-domain spectral analysis. By substituting direct time-series processing with Fourier transform-based frequency amplitude analysis, the system achieves superior sub-component failure detection capability while simplifying the computational approach through established spectral processing techniques.
Solution Approach 2:
The patent extracts specific frequency amplitude characteristics corresponding to sub-components from the comprehensive motion spectrum, focusing analysis on diagnostically relevant frequency bands rather than processing the entire frequency spectrum. This extraction targets critical failure indicators while reducing computational burden.
Data Source
AI summary
Methods and devices for diagnosing a robot. The method includes obtaining a first signal generated by a rotating component of the robot during operation of the robot. The first signal includes motion information of the rotating component. The first signal is preprocessed to filter out a part of the motion information in the first signal. The preprocessed first signal or spectrum information about the preprocessed first signal is sent to a server for diagnosing the robotU. A second signal is received from the server, wherein the second signal includes diagnostic information indicating whether a sub-component of the rotating component has a failure.