Drive Fault Detection with Speed-Normalized Vibration Spectra
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for fault recognition in drives, such as electric motors, are inadequate in accurately identifying mechanical faults and wear in components like bearings and transmissions, often relying on narrow passbands and bearing information that may be inaccurate or unavailable, leading to unreliable fault detection.
Innovation Solution
A method involving the establishment of a normalized frequency spectrum for speed and acceleration, recognizing peak values and patterns, using bandpass filters to identify specific frequency ranges, and employing artificial intelligence and expert knowledge to analyze vibration data, independent of bearing information, to detect faults and wear.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If narrow passbands and bearing information are used for fault recognition, then the method can be simpler to implement, but the accuracy and reliability of fault detection deteriorates
Solution Approach 1:
The patent transforms the vibration signal from the time domain to the frequency domain using Fast Fourier Transform (FFT), changing the parameter representation to enable more reliable fault detection. This parameter transformation allows identification of fault characteristics that are not visible in the time domain, resolving the contradiction between implementation simplicity and detection reliability.
Solution Approach 2:
The patent adds a frequency dimension to the analysis by computing the spectral content of the vibration signal. This dimensional transition from time-domain to frequency-domain enables detection of fault patterns that would be invisible in traditional time-domain analysis, improving reliability without significantly increasing implementation complexity.
2Device complexity
If traditional vibration monitoring methods are used, then the system can be simpler, but the ability to detect wear and faults in multiple components deteriorates
Solution Approach 1:
The patent creates a universal fault detection method that can identify various types of faults (bearing faults, gear faults, misalignment, etc.) and wear conditions across multiple components using a single unified approach. The frequency-domain analysis and pattern recognition algorithms can adapt to different fault types and components, providing multi-functionality without requiring separate specialized systems for each component.
Solution Approach 2:
The patent introduces the frequency spectrum as an intermediary representation that mediates between the raw vibration signal and the fault diagnosis. This spectral intermediary enables extraction of characteristic frequencies and patterns that indicate specific component faults, allowing a single system to detect multiple types of faults across different components.
3Measurement precision
If bearing information is required for fault recognition, then the method can be more accurate for bearing faults, but the method becomes less reliable when bearing information is inaccurate or unavailable
Solution Approach 1:
The patent extracts fault detection capability from the bearing information requirement. By using frequency-domain analysis and pattern recognition, the system can identify bearing faults and other component faults without requiring precise bearing parameters or information. The method extracts characteristic frequency patterns that are sufficient for fault detection independently of bearing catalog data or specifications.
Solution Approach 2:
The system performs self-service fault detection by using the vibration signal itself and its spectral characteristics to identify faults, rather than requiring external bearing information or parameters. The pattern recognition algorithms learn from the signal characteristics and can detect faults autonomously without relying on pre-stored bearing data, making the system self-sufficient.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables robust and timely fault detection in drives, reducing unplanned downtime by accurately identifying various mechanical issues, including belt misalignment and bearing wear, without reliance on precise bearing information, and facilitating predictive maintenance.
Implementation Method 1
The vibration is recorded, for example, by means of a vibration sensor, wherein the vibration is recorded, in particular, on a housing or within a transmission
Data Source
AI summary
Disclosed is a method for identifying faults in a drive. A normalized spectrum is determined, which is dependent on a speed. Peak values in the spectrum are identified. A first peak value is identified at a first frequency and a second peak value is identified at a second frequency. A pattern is identified based on the first frequency and the second frequency.