Sawtooth Wavelet Signal Extraction for Industrial IoT Noise
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Industrial systems face challenges in extracting signals from legacy machines due to strong background noise, as these machines are not designed for IIoT systems and cannot be modified, leading to low signal strength and interference from noise, making it difficult to monitor consumable tools like lathe machine inserts effectively.
Innovation Solution
A signal processing method involving a device communicatively coupled to sensing devices and a server, which converts sensor data into frequency component data, executes a correlation process to identify frequency bands correlated with a sawtooth pattern, and provides feature values with band limitation within these bands to the server, enhancing signal extraction in the presence of strong noise.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sensors are placed inside the machine enclosure to collect stronger physical signals, then signal strength is improved, but the machine modification becomes invasive and is not permitted
Solution Approach 1:
The patent uses the machine enclosure surface as an intermediary medium to transmit vibration signals from the internal consumable tools to external sensors. The enclosure acts as a mediator that allows non-invasive signal collection by transmitting mechanical vibrations through its surface, resolving the contradiction between needing strong signals and avoiding invasive modifications
Solution Approach 2:
The patent replaces direct mechanical contact sensors with acoustic/vibration sensors that detect signals through the enclosure surface. This substitution allows signal collection without physical intrusion into the machine interior, maintaining the non-invasive requirement while still capturing sufficient signal strength
2Ease of operation
If sensors are placed on the surface of the machine enclosure for non-invasive monitoring, then ease of installation is improved, but signal strength becomes weak due to propagation loss
Solution Approach 1:
The patent transforms the weak broadband vibration signal into a more detectable form by converting it to frequency component data and identifying characteristic sawtooth patterns. This parameter transformation changes the signal representation from time-domain vibration to frequency-domain characteristics, enhancing detectability despite the weak initial signal
Solution Approach 2:
The patent converts the harmful effect of noise and weak signal into a benefit by using correlation analysis to identify characteristic sawtooth patterns. The noise and weak signal characteristics are transformed into detectable pattern features through mathematical correlation, turning the disadvantage of weak signals into an advantage of pattern recognition
3Device complexity
If traditional vibration monitoring methods are used without noise filtering, then simplicity is improved, but measurement precision deteriorates due to strong background noise
Solution Approach 1:
The patent segments the broadband vibration signal into frequency components and identifies specific sawtooth-patterned frequency bands. This segmentation separates the useful signal information from the noise by dividing the frequency spectrum and isolating characteristic patterns, improving measurement precision through systematic signal decomposition
Solution Approach 2:
The patent uses correlation analysis to compare the collected vibration signal against expected sawtooth patterns, providing feedback on how well the signal matches the characteristic pattern. This feedback mechanism enables adaptive noise filtering and signal enhancement by continuously comparing actual measurements with expected patterns and adjusting the extraction process accordingly
Data Source
AI summary
Example implementations described herein are directed to systems and methods for extracting signal in presence of strong noise for industrial Internet of Things (IoT) system especially for monitoring systems of consumable items such as lathe machines, coolers and so on. Example implementations can utilize a sawtooth mother Wavelet instead of usual wavelet analysis to cleanse the incoming sensor data, thereby allowing for the converting sensor data to feature values despite having heavy noise interference.