Non-Invasive Container Fill Level Indication Using Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing non-invasive methods for measuring the fill level of carbon dioxide containers, especially those with a double-walled design, are inaccurate due to ambient noise interference and cannot be retrofitted to existing systems without causing structural issues or contaminating the contents.
Innovation Solution
A non-invasive container fill level indication system using a resonator to vibrate the outer surface, a vibration detecting device to capture response vibrations, and a machine learning algorithm to accurately determine the fill level by processing data and filtering out ambient noise, allowing for retrofitting to existing containers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a non-invasive vibration-based method is used to measure fill level, then the measurement can be performed without structural modification or contamination risk, but the measurement precision deteriorates due to ambient noise interference
Solution Approach 1:
The system uses a resonator to mechanically vibrate the container at its resonant frequency, causing the container to oscillate. A vibration sensor detects these oscillations, and the fill level is determined by analyzing the vibration characteristics. This mechanical vibration approach enables non-invasive measurement while the resonant frequency targeting improves signal strength against ambient noise.
Solution Approach 2:
The patent replaces traditional mechanical float-based level gauges with a vibration-based sensing system. Instead of using mechanical components that require installation inside the container, the system uses external vibration excitation and sensing, substituting mechanical contact methods with mechanical wave propagation through the container wall.
2Device complexity
If traditional vibration-based fill level measurement is used, then the system can operate without complex processing, but reliability deteriorates due to ambient noise interference
Solution Approach 1:
The system implements feedback by continuously monitoring the container's vibration response and using this information to determine fill level. The resonator is driven at frequencies that elicit strong resonant responses, and the feedback from the vibration sensor allows the system to track changes in resonant frequency and amplitude that correlate with fill level, improving reliability through continuous adaptive measurement.
Solution Approach 2:
The system exploits parameter changes in the container's vibrational characteristics as the fill level changes. The resonant frequency, damping ratio, and vibration amplitude all change with the amount of liquid or gas in the container. By monitoring these parameter changes rather than using a fixed threshold, the system achieves reliable measurements despite varying ambient 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
The system provides accurate, non-invasive measurement of liquid or gas volume within containers, unaffected by ambient noise, and can be retrofitted to existing systems without structural or contamination risks, ensuring reliable fill level monitoring.
Implementation Method 1
a resonator for vibrating an outer surface of the container
Implementation Method 2
a vibration detecting device for detecting a data signal indicative of a response vibration
Data Source
AI summary
A liquid container refill management system including a machine learning algorithm and method of training the same, the system and method making use of noninvasive tank-in-tank measuring techniques. The system can comprise of a container fill level indicator. The container fill level indicator can be capable of detecting a vibration response signal on the outer surface of a container, wherein the system is capable of transmitting the response signal to a remote data processor for processing using a trained machine learning algorithm. The trained machine learning algorithm can be trained by the process of selecting model inputs and outputs to define an internal structure of the machine learning algorithm, applying a collection of input and output data samples to train the machine learning algorithm, and verifying the accuracy of the machine learning algorithm by applying input data samples and comparing received output values with expected output values.