Machine-learning classification of environmental conditions inside a tank by analyzing radar curves in industrial level measurements

Borg, Denis; Sestito, Guilherme Serpa; Silva, Maíra Martins da

doi:10.1016/j.flowmeasinst.2021.101940

Cited by 7 publications

(2 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Further investigations in this direction would be of great practical interest. Here, the integration of data fusion [24] or machine learning [25][26][27] could also be of great relevance.…”

Section: Discussionmentioning

confidence: 99%

Fill level measurement of low-permittivity material using an M-sequence UWB radar

Wegner

Gebhardt²,

Galdo

2023

Int. J. Microw. Wireless Technol.

View full text Add to dashboard Cite

Due to increasingly complex and automated manufacturing processes, the demands on the control parameters of these processes are also increasing. One parameter is the fill quantity of, e.g., liquids in production plants, whose precise determination is of ever-growing importance. Up to now, the exact level of determination under difficult conditions, such as high ambient temperatures, has been a particular challenge. This paper demonstrates a novel method by which an M-sequence UWB radar can determine levels of low-permittivity materials in small metal containers. For this purpose, hot melt is used as an example. Thus, the influence of large temperature differences on the long-term stability of level measurement can also be investigated. The measurements show that the level of hot melt can be measured to be long-term stable with an accuracy of better than 3 mm. Furthermore, the precise determination of the empty state is highly important for many applications. For this reason, this paper shows a method for determining the empty state without complex calibration procedures. For the empty level indication, an accuracy of up to 0.5 mm could be achieved for molten hot glue and 3% of the tank volume, independent of the shape or aggregate state of the medium.

show abstract

“…Further investigations in this direction would be of great practical interest. Here, the integration of data fusion [24] or machine learning [25][26][27] could also be of great relevance.…”

Section: Discussionmentioning

confidence: 99%

Fill level measurement of low-permittivity material using an M-sequence UWB radar

Wegner

Gebhardt²,

Galdo

2023

Int. J. Microw. Wireless Technol.

View full text Add to dashboard Cite

show abstract

“…Categorizing documents according to domain‐specific ontologies and taxonomies to enrich information that domain experts and practitioners can analyze is a common task for working with text data (Grimmer & Stewart, 2013). In recent decades, automated categorization models have been increasingly utilized in a variety of domains such as environmental science (Borg et al, 2021) and social science (Zhang & Pan, 2019a, 2019b) to minimize the amount of labor and time needed for labeling large‐scale text data. We have access to (highly) accurate models for tasks such as positive and negative sentiment prediction (Yadav & Vishwakarma, 2020; Zhang & Pan, 2019a), latent topic discovery through topic modeling (Roberts et al, 2013), and classification of online posts with regard to collective action (Zhang & Pan, 2019a), Covid‐19 (Lu et al, 2021), and hate speech (Schmidt & Wiegand, 2017), among many others.…”

Section: Introductionmentioning

confidence: 99%

An expert‐in‐the‐loop method for domain‐specific document categorization based on small training data

Han

Rezapour

Nakamura

et al. 2022

Asso for Info Science & Tech

View full text Add to dashboard Cite

Automated text categorization methods are of broad relevance for domain experts since they free researchers and practitioners from manual labeling, save their resources (e.g., time, labor), and enrich the data with information helpful to study substantive questions. Despite a variety of newly developed categorization methods that require substantial amounts of annotated data, little is known about how to build models when (a) labeling texts with categories requires substantial domain expertise and/or in-depth reading, (b) only a few annotated documents are available for model training, and (c) no relevant computational resources, such as pretrained models, are available. In a collaboration with environmental scientists who study the socio-ecological impact of funded biodiversity conservation projects, we develop a method that integrates deep domain expertise with computational models to automatically categorize project reports based on a small sample of 93 annotated documents. Our results suggest that domain expertise can improve automated categorization and that the magnitude of these improvements is influenced by the experts' understanding of categories and their confidence in their annotation, as well as data sparsity and additional category characteristics such as the portion of exclusive keywords that can identify a category.

show abstract

Backpropagation Neural Network with Adaptive Learning Rate for Classification

Jullapak

Thammano

2023

Lecture Notes on Data Engineering and Communications Technologies

View full text Add to dashboard Cite

Machine-learning classification of environmental conditions inside a tank by analyzing radar curves in industrial level measurements

Cited by 7 publications

References 28 publications

Fill level measurement of low-permittivity material using an M-sequence UWB radar

Fill level measurement of low-permittivity material using an M-sequence UWB radar

An expert‐in‐the‐loop method for domain‐specific document categorization based on small training data

Backpropagation Neural Network with Adaptive Learning Rate for Classification

Contact Info

Product

Resources

About