Fault detection and diagnosis in electric motors using 1d convolutional neural networks with multi-channel vibration signals

Ribeiro, Ronny Francis; Methodoly, Isac Antônio dos Santos Areias; Campos, Mateus Mendes; Teixeira, Carlos Eduardo; Silva, Luiz Eduardo Borges da; Gomes, Guilherme Ferreira

doi:10.1016/j.measurement.2022.110759

Cited by 83 publications

(41 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Due to the classification of the persons’ rotation angles, the training and testing are carried out using a one-dimensional convolutional neural network to predict the actual rotation angles of the persons. The pre-treated skeleton data are featured by the one-dimensional data with a similar sound or semantic analysis [ 28 ]. Finally, the shoulder widths before the rotation of the persons are output based on the predicted angles and trigonometric function.…”

Section: Methodsmentioning

confidence: 99%

A Convolutional Neural-Network-Based Training Model to Estimate Actual Distance of Persons in Continuous Images

Tsai

Modales

Lin

2022

Sensors

View full text Add to dashboard Cite

Distance and depth detection plays a crucial role in intelligent robotics. It enables drones to understand their working environment to avoid collisions and accidents immediately and is very important in various AI applications. Image-based distance detection usually relies on the correctness of geometric information. However, the geometric features will be lost when the object is rotated or the camera lens image is distorted. This study proposes a training model based on a convolutional neural network, which uses a single-lens camera to estimate humans’ distance in continuous images. We can partially restore depth information loss using built-in camera parameters that do not require additional correction. The normalized skeleton feature unit vector has the same characteristics as time series data and can be classified very well using a 1D convolutional neural network. According to our results, the accuracy for the occluded leg image is over 90% at 2 to 3 m, 80% to 90% at 4 m, and 70% at 5 to 6 m.

show abstract

Section: Methodsmentioning

confidence: 99%

A Convolutional Neural-Network-Based Training Model to Estimate Actual Distance of Persons in Continuous Images

Tsai

Modales

Lin

2022

Sensors

View full text Add to dashboard Cite

show abstract

“…Many current methods for identifying structural damage, such as Genetic Algorithm (GA) and intelligent methods such as artificial neural networks (ANNs) are often implemented on the basis of some measured data and a large number of simulation data of structural vibration responses (Gomes et al [2019b],Ribeiro Junior et al [2020], Junior et al [2021], Junior et al [2022]). Therefore, Yan et al [2006] emphasized that the establishment of a precise and efficient dynamic model for a structure is an important precondition.…”

Section: Frequency Response Function-based Methodsmentioning

confidence: 99%

On the use of Modal Test Data in Inverse Problems: Fundamentals and Applications

Gomes¹,

Carneiro²,

S.³

et al. 2022

Fundamental Concepts and Models for the Direct Problem

View full text Add to dashboard Cite

In the past few decades, modal testing has become a major technology for determining, improving and optimizing dynamic characteristics of engineering structures in different fields, with emphasys (but not limited to) mechanical, civil, and aeronautical/aerospace engineering. One of the great advantages of using modal tests is related to their experimental practicality. This practical advantage leads to its strong use in inverse methods, or commonly known as inverse modal-based problems. An inverse problem based on modal parameters can be formulated basically using i) natural frequency, ii) damping factor or loss factor and iii) mode shapes. Other metrics can also be used from the composition of these 3 main responses, such as modal strain energy, inverse FRF and many others. Inverse modal-based identification through optimization algorithms are particularly emphasized. The methods discussed here are mainly elaborated by the evaluation of modal data due to the great potential of application. This chapter discusses the use of computational and intelligent techniques for parameter identification using modal responses. The content in this paper aims to help engineers and researchers find a starting point in developing a better solution to their specific structural problems, either by inverse methods, pattern recognition, or intelligent signal processing.

show abstract

“…The LSTM network's predicted categorization of a signal for showing the presence of damage. Using two accelerometers measuring in two different directions, Junior et al (2022) suggested a multi‐head 1D CNN to identify and diagnose six distinct sorts of problems in an electric motor. A series of tests with seven distinct induced faults and operation conditions are used to validate the suggested approach.…”

Section: Applications Of ML In Solving Dynamical Problemsmentioning

confidence: 99%

Machine intelligence in dynamical systems: \A state‐of‐art review

Sahoo

Chakraverty

2022

WIREs Data Min & Knowl

View full text Add to dashboard Cite

This article is dedicated to study the impact of machine intelligence (MI) methods viz. various types of Neural models for investigating dynamical systems arising in interdisciplinary areas. Different types of artificial neural network (ANN) methods, viz., recurrent neural network, functional‐link neural network, convolutional neural network, symplectic artificial neural network, genetic algorithm neural network, and so on, are addressed by different researchers to investigate these problems. Although various traditional methods have been developed by researchers to solve these dynamical problems but the existing traditional methods may sometimes be problem dependent, require repetitions of the simulations, and fail to solve nonlinearity behavior. In this regard, neural network model based methods are more general and solutions are continuous over the given domain of integration, self‐adaptive and can be used as a black box. As such, in this article, we have reviewed and analyzed different MI methods, which are applied to investigate these problems. This article is categorized under: Technologies > Computational Intelligence Technologies > Machine Learning Application Areas > Science and Technology

show abstract

Fault detection and diagnosis in electric motors using 1d convolutional neural networks with multi-channel vibration signals

Cited by 83 publications

References 32 publications

A Convolutional Neural-Network-Based Training Model to Estimate Actual Distance of Persons in Continuous Images

A Convolutional Neural-Network-Based Training Model to Estimate Actual Distance of Persons in Continuous Images

On the use of Modal Test Data in Inverse Problems: Fundamentals and Applications

Machine intelligence in dynamical systems: \A state‐of‐art review

Contact Info

Product

Resources

About