Automatic parametric fault detection in complex analog systems based on a method of minimum node selection

Bilski, Adrian; Wojciechowski, Jacek

doi:10.1515/amcs-2016-0045

Cited by 10 publications

(6 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Note that similar analysis must be performed for every circuit separately. For larger circuits, automated node selection methods are proposed [35]. The presented experiments show usefulness of the RBF network as the parametric fault classifier.…”

Section: Experimental Examplementioning

confidence: 91%

Diagnostics of analog systems using Artificial Neural Networks

Bilski¹

2017

PAR

View full text Add to dashboard Cite

ŎƬ˘˹ŘǚŘ ʊǔǄ ȭŘ ǘɁʁ˘˿ʊǜŘȭǔƬ ˘ Řʁǜ˿ǘˁȦˁ ȭŘ ˹ŘʁˁȭǘŘƋǒ ǚǔƋƬȭƋǖǔ ,ʁƬŘǜǔ˸Ƭ ,ɁȧȧɁȭʊ Ę˘ȭŘȭǔƬ ŘˁǜɁʁʊǜ˹Ř ʯƖ1 . IntroductionDiagnostics of analog systems is currently a wide domain with numerous applications, closely related to the types of analyzed objects. With the increasing number of such systems (including mechanical and electronic elements), the set of usable approaches also increases. Contrary to the digital object, its analog counterpart is characterized by the continuous transmission and processing of information. Therefore the system may also be in the infinite number of states, generating continuous signals. This makes the diagnostics of analog systems especially difficult. The most popular approaches used in the domain belong to Artificial Intelligence (AI). Their advantages include the autonomous operation (without the input of the human operator), high accuracy with generalization abilities and (in most cases) the ability to update knowledge through the additional learning procedure. Disadvantages include the dependency of the acquired diagnostic knowledge on the available data (use cases) presented during the training of the diagnostic module. Sometimes explanation of the reasoning process (i.e. how the premises were used to draw conclusions) is required. The most popular diagnostic methods, exploited in practice, are Artificial Neural Networks (ANN) [1,2]. This paper presents the taxonomy, structural and operational details of ANN used for the diagnostic purposes. The structure of the paper is as follows. Section 2 presents diagnostic principles important from the AI-application point of view. In Section 3 the generic diagnostic architecture is introduced with the focus on internal modules. Section 4 covers the ANN work regime from the diagnostic point of view. In Section 5 the main architectures of ANN with their applications are presented. Section 6 presents diagnostics of the 5 th order lowpass filter using the RBF ANN-based classifier. In Section 7 summary and future prospects of the discussed ANN are provided. Aims of the diagnosticsDiagnostics is aimed at determining the actual state of the analyzed System Under Test -SUT, based on the observations of the measured signals y(t) being responses to the known excitation x(t). Similarly to the medical and veterinary diagnostics, in technical sciences non-invasive approaches are preferred, where the information about the system's behavior is available only through the input and output (accessible and partially accessible) nodes (with the maximum set of nodes being hidden, i.e. inaccessible). Therefore the diagnostic system is expected to establish relations (in the form of the unknown function f()) between the observable features of output signals and configuration of parameters p defining the SUT's work regime [1]. The latter can be based on the human expert's experience or the mathematical description of the system. Alternatively, simulations of the SUT model may provide the information about how it works and how changes in parameters influence behavior o...

show abstract

Section: Experimental Examplementioning

confidence: 91%

Diagnostics of analog systems using Artificial Neural Networks

Bilski¹

2017

PAR

View full text Add to dashboard Cite

show abstract

“…Support vector machines. Support vector machines (SVMs), which often belong to the most effective general-purpose classification algorithms (e.g., Hamel, 2009;Cichosz, 2015;Bilski and Wojciechowski, 2016), can be viewed as a considerably strengthened version of a basic linear-threshold classifier with the following enhancements (Cortes and Vapnik, 1995;Platt, 1998): margin maximization: the location of the decision boundary (separating hyperplane) is optimized with respect to the classification margin;…”

Section: 3mentioning

confidence: 99%

A Case Study in Text Mining of Discussion Forum Posts: Classification with Bag of Words and Global Vectors

Cichosz

2018

International Journal of Applied Mathematics and Computer Science

View full text Add to dashboard Cite

Despite the rapid growth of other types of social media, Internet discussion forums remain a highly popular communication channel and a useful source of text data for analyzing user interests and sentiments. Being suited to richer, deeper, and longer discussions than microblogging services, they particularly well reflect topics of long-term, persisting involvement and areas of specialized knowledge or experience. Discovering and characterizing such topics and areas by text mining algorithms is therefore an interesting and useful research direction. This work presents a case study in which selected classification algorithms are applied to posts from a Polish discussion forum devoted to psychoactive substances received from home-grown plants, such as hashish or marijuana. The utility of two different vector text representations is examined: the simple bag of words representation and the more refined embedded global vectors one. While the former is found to work well for the multinomial naive Bayes algorithm, the latter turns out more useful for other classification algorithms: logistic regression, SVMs, and random forests. The obtained results suggest that post-classification can be applied for measuring publication intensity of particular topics and, in the case of forums related to psychoactive substances, for monitoring the risk of drug-related crime.

show abstract

“…The procedure for other similar circuits (such as other filters) should be the same. If the system is more complex, more sophisticated node selection method should be used [27].…”

Section: Testability Verificationmentioning

confidence: 99%

Analysis of the RBF ANN-based classifier for the diagnostics of electronic circuit

Bilski

Połok

2018

ACTA IMEKO

View full text Add to dashboard Cite

<span lang="EN-GB">The paper presents the application of the Radial Basis Function (RBF) Artificial Neural Network (ANN) to the diagnostics of the analog circuit. Such networks are in most cases useful in the approximation tasks as the alternative to multilayered perceptrons (MLP) or Support Vector Machines (SVM). In this work the analysis of various RBF ANN-based classifier configurations for the fault detection and identification module are is conducted. The considered parameters included the optimal number of neurons in the hidden layer, coding schemes for the output layer neurons and operation duration during the training and testing the classifier. The efficiency of the diagnostic system is verified using the fifth order lowpass filter. The circuit was also analyzed in terms of the testability, depending on the set of accessible nodes, confronted against the output node only. Experiments cover also accuracy comparison between the RBF, MLP and SVM classifiers. Results show advantages and drawbacks of the RBF ANN-based diagnostic module, compared to other available solutions.</span>

show abstract

Automatic parametric fault detection in complex analog systems based on a method of minimum node selection

Cited by 10 publications

References 33 publications

Diagnostics of analog systems using Artificial Neural Networks

Diagnostics of analog systems using Artificial Neural Networks

A Case Study in Text Mining of Discussion Forum Posts: Classification with Bag of Words and Global Vectors

Analysis of the RBF ANN-based classifier for the diagnostics of electronic circuit

Contact Info

Product

Resources

About