Machine Acoustic Reference Signals for Low-Data Condition Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing acoustic analysis methods for machine condition monitoring require large amounts of data, high computational effort, and significant instrumentation costs, with limited scope of application and lack of automation.
Innovation Solution
Generate individual acoustic signals for each machine component, determine an expected overall acoustic signal based on these, and create a reference signal using a computing unit, allowing for component-specific analysis without the need to record individual operating noises of all components.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If process-acquired state data from a machine is analyzed using machine learning techniques to derive meaningful trends, then reliable predictions can be made, but large amounts of data and correspondingly high computing power are required
Solution Approach 1:
The patent extracts only the relevant acoustic information from machine operation by comparing actual acoustic signals against a pre-established reference signal containing normal operation characteristics. This extraction approach eliminates the need to process large volumes of raw data while maintaining prediction reliability, as only deviations from the reference are analyzed.
Solution Approach 2:
The patent performs preliminary action by pre-acquiring and storing reference acoustic signals during normal machine operation before actual condition monitoring begins. This reference data is processed beforehand to establish baseline characteristics, enabling subsequent real-time analysis to focus only on detecting deviations rather than processing all raw operational data.
2Reliability
If process-acquired state data from a machine is analyzed using machine learning techniques to derive meaningful trends, then reliable predictions can be made, but correspondingly high computing power is required
Solution Approach 1:
The patent extracts only the relevant acoustic information from machine operation by comparing actual acoustic signals against a pre-established reference signal containing normal operation characteristics. This extraction approach eliminates the need to process large volumes of raw data while maintaining prediction reliability, as only deviations from the reference are analyzed.
Solution Approach 2:
The patent performs preliminary action by pre-acquiring and storing reference acoustic signals during normal machine operation before actual condition monitoring begins. This reference data is processed beforehand to establish baseline characteristics, enabling subsequent real-time analysis to focus only on detecting deviations rather than processing all raw operational data.
3Reliability
If acoustic analysis is performed on machine operating noise, then machine state can be monitored, but large instrumentation costs are required
Solution Approach 1:
The patent applies universality by using a single acoustic sensor to perform multiple functions: it captures reference acoustic signals during normal operation, monitors operational acoustic signals during actual use, and enables condition monitoring through comparison analysis. This multi-functional approach eliminates the need for separate instrumentation systems for different monitoring purposes.
Solution Approach 2:
The patent creates a acoustic copy of normal machine operation by recording and storing reference signals that represent healthy machine states. This acoustic replica serves as a template for comparison against actual operational signals, enabling condition monitoring without requiring complex reference instrumentation systems.
4Reliability
If acoustic analysis is performed on machine operating noise, then machine state can be monitored, but limited scope of application results
Solution Approach 1:
The patent applies universality by using a single acoustic sensor to perform multiple functions: it captures reference acoustic signals during normal operation, monitors operational acoustic signals during actual use, and enables condition monitoring through comparison analysis. This multi-functional approach eliminates the need for separate instrumentation systems for different monitoring purposes.
Solution Approach 2:
The patent enables broad application scope by adapting to different machine types and operating conditions through parameter changes in the reference signal rather than through hardware modifications. The system can monitor various machines by establishing appropriate reference acoustic profiles for each specific application, making the monitoring approach universally applicable across different contexts.
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
Reduces data and computational requirements, minimizes instrumentation, and enables automated, universal applicability for machine condition analysis, allowing on-premises analysis without complex data processing.
Implementation Method 1
a sound transducer configured to generate individual acoustic signals depending on individual operating noises of at least two machine components
Data Source
Figure 1
Figure 2~3
AI summary
In order to provide an acoustic analysis of a condition of a machine (8) having reduced computational complexity, a method for generating at least one reference signal (20) is provided. An individual acoustic signal is provided for each of at least two components (14, 15, 16) of the machine (8). A total acoustic signal which is to be expected is determined on the basis of the individual signals by means of a computing unit (13). The reference signal (20) is generated by means of the computing unit (13) in accordance with the total signal.