Fault Diagnosis Model Using Stacked RBMs for Vibration Signals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional fault diagnosis methods for mechanical devices operating in harsh environments, such as high magnetic fields or high temperatures, are inefficient and inaccurate due to the difficulty in extracting characteristic frequencies from vibration signals, especially for devices in concealed spaces.
Innovation Solution
A system utilizing a fault diagnosis model comprising a trained first component with stacked Restrictive Boltzmann Machines (RBMs) and a fully connected layer, which processes vibration signal features like Kurtosis and Skewness to determine fault conditions automatically, improving accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional vibration signal analysis methods are used, then fault diagnosis can be performed, but the diagnosis accuracy and efficiency deteriorate due to difficulty in extracting characteristic frequencies
Solution Approach 1:
The patent replaces conventional mechanical signal processing methods (Fast Fourier Transformation, spectral envelope analysis) with a machine learning-based fault diagnosis model. This model automatically learns and extracts characteristic frequencies from vibration signals without requiring manual signal processing operations, thereby improving both diagnosis accuracy and efficiency simultaneously.
Solution Approach 2:
The fault diagnosis model performs self-learning and automatic feature extraction from vibration signals. It independently identifies characteristic frequencies and fault patterns without requiring manual intervention or complex preprocessing steps, enabling the system to serve itself in extracting diagnostic features while improving efficiency and accuracy.
2Ease of operation
If manual fault diagnosis is performed for devices in concealed spaces, then some fault information can be obtained, but the diagnosis accuracy deteriorates due to inability to access devices
Solution Approach 1:
The patent introduces vibration sensors as intermediaries that can be attached to externally accessible surfaces of devices in concealed spaces. These sensors capture vibration signals that propagate through the device structure, enabling remote fault diagnosis without direct access to internal components. The fault diagnosis model then processes these signals to achieve accurate fault detection despite the devices being in concealed locations.
3Measurement precision
If additional signal processing operations are performed to extract characteristic frequencies, then fault diagnosis can be achieved, but the processing time and complexity increase
Solution Approach 1:
The fault diagnosis model performs preliminary learning and feature extraction during the training phase, automatically identifying and memorizing characteristic frequency patterns associated with different fault types. During actual diagnosis, the model directly applies this pre-learned knowledge to quickly identify faults without requiring time-consuming signal processing operations like Fast Fourier Transformation or spectral envelope analysis, significantly reducing processing time while maintaining accuracy.
Data Source
AI summary
A system for fault diagnosis is provided. The system may acquire a vibration signal of a target device, and determine one or more feature values of the vibration signal. The system may further determine a fault condition of the target device by applying a fault diagnosis model to the feature values. The fault diagnosis model may include a trained first component including a plurality of stacked trained RBMs, and a trained second component connected to the trained first component. The trained second component may include a trained fully connected layer and a trained output layer connected to the trained fully connected layer.