Multi-Sensor Signal Fusion for Non-Stationary Machine Health Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional machine monitoring and diagnostic techniques are poorly suited for non-stationary machines due to the time-varying nature of their operational characteristics, leading to inaccurate analysis of signals such as vibration and magnetic flux, which are crucial for assessing machine health and performance.
Innovation Solution
A system utilizing multiple sensors to acquire non-stationary signals, fuse these signals using wavelet transforms or deep learning, and extract features insensitive to the machine's operational stationarity, enabling accurate analysis and predictive maintenance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional machine monitoring techniques are used on non-stationary machines, then the system complexity remains low, but the measurement precision and reliability of machine condition assessment deteriorate due to non-stationary signal characteristics
Solution Approach 1:
The patent segments non-stationary signals into multiple stationary sub-signals using signal decomposition techniques (e.g., wavelet transform, empirical mode decomposition). This allows conventional analysis methods to be applied to each stationary component, improving measurement precision while managing system complexity through modular processing
Solution Approach 2:
The patent employs dynamic signal processing methods that adapt to changing signal characteristics over time. By using time-frequency analysis and adaptive filtering techniques, the system maintains high measurement precision for non-stationary signals without requiring overly complex fixed-structure processors
2Reliability
If multiple sensors are used to acquire non-stationary signals, then the measurement precision and reliability improve, but the device complexity and data processing requirements increase
Solution Approach 1:
The patent merges data from multiple sensors through signal fusion techniques, combining complementary information to improve reliability of machine health assessment. The fusion process integrates vibrations, acoustic emissions, and other sensor data to create a more robust diagnostic picture
Solution Approach 2:
The patent introduces intermediary processing layers including feature extraction modules and data fusion algorithms that mediate between raw sensor data and final diagnostic conclusions. These intermediaries simplify the integration of multiple sensor inputs while maintaining high reliability
3Manufacturing precision
If feature extraction methods sensitive to stationarity are used, then the manufacturing precision of feature extraction is high for stationary signals, but the reliability deteriorates when applied to non-stationary machine operation
Solution Approach 1:
The patent develops dynamic feature extraction methods that adapt to non-stationary conditions. By using time-varying statistical parameters and adaptive thresholding, the system maintains high extraction accuracy across changing operational conditions rather than relying on fixed stationary assumptions
Solution Approach 2:
The patent changes the parameters used for feature extraction to be insensitive to stationarity. By selecting and emphasizing certain signal characteristics that remain stable under non-stationary operation, the system maintains both precision and reliability across varying machine conditions
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 highly accurate analysis of non-stationary machine performance, allowing for timely maintenance and optimization of operational parameters by extracting features that are insensitive to the machine's non-stationary operation, thereby improving reliability and efficiency.
Implementation Method 1
the fusing includes applying a wavelet transform to the first and second non-stationary outputs and the modifying includes multiplying the wavelet transform of one of the first and second non-stationary outputs by a binary mask of the wavelet transform of the other one of the first and second non-stationary outputs
Data Source
AI summary
A method for monitoring at least one machine including causing at least a first sensor to acquire at least a first non-stationary signal from at least one machine operating in a non-stationary manner during at least one operational time frame, the first sensor(s) providing at least a first non-stationary output, causing at least a second sensor to acquire at least a second non-stationary signal from the machine(s) during the operational time frame(s), the second sensor(s) providing at least a second non-stationary output, fusing the first non-stationary output(s) with the second non-stationary output(s) to produce a fused output, extracting at least one feature from the first and/or second non-stationary signal(s) based on the fused output, analyzing the feature(s) to ascertain a state of health of the machine(s) and performing a repair operation, maintenance operation and/or modification of operating parameters of the machine(s) based on the analyzed state of health.