Advanced searches left 3/3

FAULT - PubMed

Summarized by Plex Scholar
Last Updated: 24 September 2022

* If you want to update the article please login/register

Data-driven fault detection and isolation of nonlinear systems using deep learning for Koopman operator.

This paper introduces a data-driven actuator fault detection and isolation scheme for the general class of nonlinear systems. The proposed method uses a deep neural network architecture to obtain an invariant set of base functions for the Koopman operator, resulting in a linear Koopman predictor for a nonlinear system. The linear model developed in Then, the resulting linear model is used for fault detection and isolation purposes without relying on prior knowledge of the underlying dynamics.

Source link: https://doi.org/10.1016/j.isatra.2022.08.030


Machine fault detection methods based on machine learning algorithms: A review.

Preventive identification of mechanical parts failures has always played a vital role in machine repair. With the processing cycles being repeated, the machinery in the production process is subjected to wear over time, resulting in a loss of technical reliability in comparison to optimal conditions. For this reason, it is imperative to plan failures in advance in order to replace the part before its wear causes a decrease in machine results. Several systems have recently been introduced for the prevention of fault detection by using a combination of low-cost sensors and algorithms based on machine learning. Support Vector Machine solutions, Artificial Neural Network algorithms, Convolutional Neural Network modeling, Recurrent Neural Network solutions, and Deep Generative Systems are among the most commonly used machine learning algorithms investigated in this research. The most common mechanical failures are identified in this research, as well as machine learning algorithms, are investigated.

Source link: https://doi.org/10.3934/mbe.2022534


Detraque: Dynamic execution tracing techniques for automatic fault localization of hardware design code.

To find mistakes, the most common Verilog fault localization schemes use the static analysis technique. Detraque traces these executions to localize errors in test cases, so the likelihood of any Verilog statement being inaccurate and sorting the statements in decreasing order by suspicion rank can be determined. Detraque can achieve a score of 18. 3% thanks to empirical investigation into real Verilog programs with 61 flawed versions. Since used as a complement to static analysis techniques, Detraque has been shown to be able to increase Verilog fault localization efficiencies.

Source link: https://doi.org/10.1371/journal.pone.0274515


Application of CNN-Based Machine Learning in the Study of Motor Fault Diagnosis.

People are in a critical position in the manufacturing and life of people, with the growth of science and technology, rapid growth of social economy, and the introduction of the motor as a new form of transmission technology. Manufacturing machinery is getting more accurate, faster, more consistent, and more automated as a result of the rapid growth of computer and electronics technology. A CNN-based machine learning fault diagnosis scheme is suggested in this paper to solve the issue of high incidence of motor faults and difficulty in determining fault types. To extract vibration time domain data for normal operating conditions, rotor eccentricity, stator short circuit, and bearing inner ring fault, machine learning techniques are used to extract vibration time domain data; divide the data segment into 15 speed segments; and perform numerical analyses for fault diagnosis.

Source link: https://doi.org/10.1155/2022/9635251


Bioinspired Injectable Self-Healing Hydrogel Sealant with Fault-Tolerant and Repeated Thermo-Responsive Adhesion for Sutureless Post-Wound-Closure and Wound Healing.

The dynamic hydrogel cross-linked through Schiff base bond, catechol-Fe coordinate bond, and the close cooperation between GT with temperature-dependent phase transition and SA endows the resulting hydrogel with the correct mechanical and adhesive strength for effective wound closure, injectionability, and self-healing capacity, as well as repeated closure of reopened wounds. The in vivo closure study revealed their ability to promote wound closure and wound healing of the incisions, which suggests that the reversible adhesive hydrogel dressing could be a versatile tissue sealant.

Source link: https://doi.org/10.1007/s40820-022-00928-z


Fault-tolerant control design for unreliable networked control systems via constrained model predictive control.

This paper addresses the challenge of passive fault-tolerant control of discrete-time networked control systems. Network flaws, such as random time delay and packet dropout, have been modeled as a Markov chain that results in a Markovian jump linear device. Some of the elements in the transition probability matrix are expected to be unknown in order to solve network complexities. A constrained model predictive control system is recommended to design a fault-tolerant control scheme in which all these topics are considered as well as input limitation.

Source link: https://doi.org/10.1016/j.isatra.2022.08.019


Efficient and Fast Joint Sparse Constrained Canonical Correlation Analysis for Fault Detection.

The canonical correlation study has drew a lot of attention in fault finding. To improve detection results, we recommend a new joint sparse constrained CCA model that incorporates the l2,0-norm joint sparse constraints into classical CCA. JSCCCA's main argument is that the number of extracted variables can be determined by using the joint sparse system to determine the number of extracted variables. Using the improved iterative hard thresholding and manifold constrained gradient descent technique, we then create an effective alternating minimization scheme.

Source link: https://doi.org/10.1109/TNNLS.2022.3201881


Variational Attention-Based Interpretable Transformer Network for Rotary Machine Fault Diagnosis.

Deep learning technologies can be a valuable tool for rotary machine fault diagnosis, where vibration signals are often used as part of a deep network model to determine the machine's internal state. Transformer networks are capable of capturing association links via the global self-attention system in order to enhance vibration signal representations. Despite this, transformer networks are unable to establish a causal relationship between signal patterns and fault types, resulting in poor interpretability. The variational attention-based transformer network, which is an interpretable deep learning framework, is designed for RMFD to solve these challenges. Two experimental studies, along with a heat map of attention weights, reveal the causal relationship between fault types and signal patterns.

Source link: https://doi.org/10.1109/TNNLS.2022.3202234


Fault Tolerance Optimization of a Lithium Battery Pack Having a Damaged Unit.

Electric vehicles are often praised for solving the most common issues of energy and air pollution as a type of green and renewable technology. In this paper, fault tolerance enhancement of an air-cooled lithium battery pack carrying a damaged unit was considered to enhance heat dissipation's heat dissipation results. A quadratic polynomial response function was established for establishing the relationship between the objective function and the design variables, which were based on the experimental findings, which were chosen by the Latin-hypercube design of experimentation's experimentation method.

Source link: https://doi.org/10.1021/acsomega.2c03329


Preliminary Assessment of the Safety of a Fault-Tolerant Control-based Wearable Tremor Suppression Glove.

The emergence of wearable tremor suppression de-vices has provided a promising alternative to parkinsonian tremor control, especially for people whose tremors are not addressed by traditional medical methods. Currently, WTSDs research has produced promising results with a tremor suppression rate of up to 99 percent; however, user security of WTSDs has not been properly considered, particularly in the occurrence of unexpected events, such as faults and disruptions, which have not been adequately considered. For the first time, a fault-tolerant control system was developed and integrated into a WTSD's control scheme. On 18 tremor motion datasets, including determining the tremor suppression ratio and error when tracking voluntary motion, were tested and evaluated, including the tremor suppression ratio and error when monitoring voluntary movement.

Source link: https://doi.org/10.1109/EMBC48229.2022.9871546

* Please keep in mind that all text is summarized by machine, we do not bear any responsibility, and you should always check original source before taking any actions

* Please keep in mind that all text is summarized by machine, we do not bear any responsibility, and you should always check original source before taking any actions