Vibration Analysis Using Multidimensional Pattern Databases
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for monitoring machining processes and other industrial operations like welding, forming, and component testing are inadequate, as they rely on limited vibration analysis techniques that fail to provide reliable quality assessment and often require costly and invasive testing for safety-critical applications.
Innovation Solution
A method for multidimensional vibration analysis using a sound sensor to record and evaluate vibrations across different frequencies and times, enabling real-time assessment of components, tools, and machining processes through pattern recognition and data comparison against reference patterns, allowing for precise identification of errors and quality evaluation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If low-frequency vibration monitoring is used to detect imbalances and tool vibrations, then rotating component faults can be identified, but machining process quality evaluation is only conditionally suitable and inadequate
Solution Approach 1:
The patent transitions from one-dimensional low-frequency vibration amplitude monitoring to three-dimensional high-frequency vibration spectrum analysis (frequency, time, amplitude dimensions). This dimensional expansion enables comprehensive capture of machining process characteristics, allowing both fault detection and precise quality evaluation to be achieved simultaneously through multidimensional pattern recognition.
2Measurement precision
If high-frequency sound signals are recorded during machining, then chatter vibrations can be detected, but the frequency range between 20 kHz and 2 MHz cannot be effectively transmitted or recognized via feeler arm coupling
Solution Approach 1:
The patent replaces mechanical feeler arm coupling with direct piezoelectric sensor coupling to the tool. This substitution eliminates the transmission limitations of mechanical coupling, enabling effective detection of high-frequency vibrations up to 2 MHz without signal loss, while simplifying the overall measurement system architecture.
3Measurement precision
If optical systems are used to monitor laser processes by measuring reflected light, then surface absorption can be assessed, but welding through quality and thermal penetration cannot be reliably checked
Solution Approach 1:
The patent introduces acoustic emission signals as an intermediary that penetrates through the material to provide information about internal welding quality and thermal penetration. While optical systems measure only surface properties, the acoustic waves propagate through the welded joints, enabling reliable assessment of subsurface welding quality and complete fusion without requiring direct visual access to the weld zone.
4Ease of operation
If conventional vibration analysis methods are used for component testing, then basic operating states can be monitored, but reliable detection of component failures during operation is not achieved
Solution Approach 1:
The patent performs preliminary action by establishing comprehensive reference vibration patterns for both healthy and faulty component states before operational monitoring begins. During operation, real-time vibration signals are continuously compared against these pre-established references using pattern recognition algorithms, enabling reliable detection of component failures as they develop, rather than waiting for obvious symptoms to appear.
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
This approach provides a reliable, real-time evaluation of machining quality and component health, enabling automated monitoring and quality assurance across various processes, including machining, welding, and component testing, reducing the need for invasive testing and improving safety in critical applications.
Implementation Method 1
The sensor (2) is preferably a structure-borne noise sensor, e.g. a piezo sensor
Data Source
Figure 1~2
Figure 3~4
Figure 5
AI summary
The invention relates to a method and a device for sound emission analysis, wherein vibrations arising during the use of a component, during the testing of a component or during the processing of a workpiece by machining, welding, forming, joining and/or cutting or the like are detected and evaluated, wherein a vibration spectrum is detected at different times or continuously and subjected to a multidimensional evaluation.