Compressed Sensor Monitoring for Low-Bandwidth Fault Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sensor-based 'big data' applications in manufacturing are inefficient and uneconomical due to the high volume of data generated, requiring significant resources for transmission, storage, and processing.
Innovation Solution
Implementing a compressed sensing technique to undersample sensor data, allowing for efficient data transmission and analysis by leveraging sparse structures in underlying signals, reducing the amount of data needed for fault detection and prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If full sensor data is collected and transmitted, then measurement precision is improved, but data transmission cost and processing resources increase significantly
Solution Approach 1:
The patent extracts only the essential features and characteristics from the full sensor data using compressed sensing techniques. Instead of transmitting all raw sensor data, the system extracts a compressed representation that contains the critical information needed for fault detection, thereby reducing transmission costs while maintaining detection accuracy.
Solution Approach 2:
The patent transforms the data representation parameters by applying compressed sensing algorithms that change the data from a high-dimensional raw format to a low-dimensional compressed format. This parameter transformation allows the essential fault detection information to be preserved while significantly reducing the data volume for transmission.
2Loss of information
If full sensor data is stored, then data completeness is improved, but storage requirements and costs increase
Solution Approach 1:
The patent extracts the essential information content from the full sensor data and stores only the compressed representation. This extraction process removes redundant data while preserving the critical fault detection information, thereby reducing storage requirements without significant loss of diagnostic capability.
Solution Approach 2:
The patent performs preliminary compression of the sensor data at the source before transmission and storage. By applying compressed sensing algorithms locally, the system prepares the data in a compact form that reduces both transmission and storage requirements while maintaining the essential information for later analysis.
3Measurement precision
If full sensor data is processed, then fault detection accuracy is improved, but processing time and computational resources increase
Solution Approach 1:
The patent extracts the essential fault detection features from the full sensor data through compressed sensing, creating a reduced dataset that retains the critical information. This extraction allows processing to focus only on the relevant features, significantly reducing computational time and resources while maintaining detection accuracy.
Solution Approach 2:
The patent applies partial processing by working with a compressed subset of the data rather than the complete dataset. The compressed sensing technique processes only the essential components of the signal, achieving sufficient fault detection accuracy with reduced computational effort and faster processing times.
4Loss of energy
If data is heavily undersampled, then data transmission volume is reduced, but signal reconstruction accuracy deteriorates
Solution Approach 1:
The patent applies parameter changes through compressed sensing algorithms that transform the undersampled data into a reconstructed signal. By changing the mathematical representation and applying appropriate reconstruction algorithms, the system recovers the essential signal characteristics from the undersampled data, maintaining acceptable reconstruction accuracy while achieving significant transmission volume reduction.
Data Source
AI summary
A method for machine monitoring is disclosed. The method includes obtaining a subset of the plurality of first data points from a local sensor connected to a machine; generating a plurality of second data points based on the subset of the plurality of first data points based on a compressed sensing technique; and identifying at least one of an operating fault or a prediction of an operating fault, based on the plurality of second data and communicating the same to an operator of the machine. Further, the method may exploit machine learning based on the low-dimensional representation learned using undersampled data. The amount of undersampling can be governed by the information content in the signal. Other aspects, embodiments, and features are also claimed and described.