Radar Boundary Layer Detection Using ML Signal Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing radar-based level measurement systems struggle to reliably detect and determine the presence and thickness of boundary layers formed by impurities or additives in the fill material, as these layers often do not generate a clear signal maximum, disrupting the measurement of the actual fill level.
Innovation Solution
A radar-based measuring system incorporating a machine learning algorithm to analyze the received signal for determining the mass or volume fraction of the boundary layer, using a TDR method with a measuring probe, which applies a machine learning algorithm to detect the distribution of the attenuation coefficient and dielectric value along the probe.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If radar-based measurement methods are used to determine fill level, then continuous and contactless measurement is achieved, but reliable detection of boundary layers is lost due to weak or absent signal maxima
Solution Approach 1:
The system performs preliminary classification of received signals into single reflection signals and multiple reflection signals before processing. This preliminary action allows the evaluation unit to apply different processing strategies to different signal types, improving boundary layer detection by preventing multiple reflection artifacts from masking the boundary layer signal maximum.
Solution Approach 2:
The patent introduces an intermediary classification mechanism that separates single reflection signals from multiple reflection signals. This intermediary step enables the system to handle each signal type appropriately, with multiple reflection signals being excluded or downweighted in the evaluation, thereby preventing them from interfering with boundary layer detection.
2Adaptability or versatility
If additive layers such as foam, sediment, or phase separations form in the fill material, then measurement disruption occurs, but the ability to detect and characterize these layers is limited by conventional signal maximum identification
Solution Approach 1:
The system classifies received signals into single reflection and multiple reflection categories before processing. This preliminary classification enables the evaluation unit to focus on single reflection signals that contain authentic boundary layer information, while excluding multiple reflection signals that would otherwise contaminate the analysis and reduce detection versatility.
Solution Approach 2:
The patent extracts and isolates single reflection signals from the mixed signal pool by classifying and separating them from multiple reflection signals. This extraction process concentrates the analysis on the most reliable signal type for boundary layer detection, improving both the precision and versatility of additive layer characterization.
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 effectively identifies and quantifies the boundary layer's parameters, such as thickness and composition, enhancing the accuracy of fill level measurements by leveraging machine learning to interpret complex signal reflections.
Implementation Method 1
radar-based measurement methods have become established for measuring the fill level of contents in containers
Implementation Method 2
the high-frequency signal is reflected in the measuring probe at the level of the contents and guided back along the probe
Implementation Method 3
TDR (Time Domain Reflectometry) - An advantage of this variant is that, due to the guided signal transmission, less power is required to operate the level gauge
Data Source
Figure 1~2
AI summary
The invention relates to a measuring system for measuring a boundary layer of a filler (2, 2') in a container (3). The measuring system comprises the following system components: a radar-based measuring device (1) comprising an antenna or a measuring sensor (11), by means of which high-frequency signals (SHF) can be transmitted towards the filler (2, 2') and can be received as received signals (RHF) after being reflected on the filler surface; a signal generating unit which generates the high-frequency signal (SHF) to be transmitted; and a receiving unit which is designed to record the received signal (RHF). According to the invention, the measuring system additionally comprises an analysis unit (4) in which a machine learning algorithm (MLA), such as an artificial neural network, is designed to detect the boundary layer using the received signal (RHF). By implementing a machine learning algorithm (MLA) according to the invention, the invention solves the problem of the boundary layer or the additive filler layer (2') often not generating a uniquely assignable signal maximum in the received signal (RHF), as is required in known clearance measuring methods according to the prior art in order to detect a clearance.