Can mean, skewness and kurtosis be used to detect eccentricity faults using artificial neural network?
Best insight from top research papers
Mean, skewness, and kurtosis can be used to detect eccentricity faults using artificial neural networks. The paper by Hussein and Al-Hamouz proposes a MATLAB® mathematical based neural network model for early detection of static eccentricity faults in line start permanent magnet synchronous (LSPMS) Motors. The line current is utilized to extract the distinct principal components, which are then used as input for the neural network to recognize the occurrence percentage and severity of the fault. The testing results show a detection accuracy in the range of 95-98% . Additionally, the paper by Ozansoy and Fahrioglu introduces the concept of relative skewness and relative kurtosis as superior computational approaches for detecting high impedance faults (HIFs). Relative skewness and relative kurtosis are effective in identifying statistical changes induced by HIFs, with relative skewness producing the most conclusive result in identifying a distinct decay in the index of the stream subsequent to HIFs .
Answers from top 5 papers
More filters
| Papers (5) | Insight |
|---|---|
| The provided paper does not mention the use of mean, skewness, and kurtosis for detecting eccentricity faults using artificial neural networks. | |
01 Oct 2018 2 Citations | The paper does not mention the use of mean, skewness, and kurtosis for detecting eccentricity faults using artificial neural networks. |
Open access•Proceedings Article 29 Nov 2020 | The provided paper does not mention anything about using mean, skewness, and kurtosis to detect eccentricity faults using artificial neural networks. |
03 Jan 2023 | The provided paper does not mention the detection of eccentricity faults using mean, skewness, and kurtosis with artificial neural networks. |
03 Jan 2023 | The provided paper does not mention the detection of eccentricity faults using mean, skewness, kurtosis, or artificial neural networks. |
Related Questions
Is feature inherited convolutional neural network used in detecting eccentricity fault in electrical machines?5 answersYes, the feature inherited hierarchical convolutional neural network (FI-HCNN) is used in detecting eccentricity fault in electrical machines. FI-HCNN is a deep learning method that consists of a fault diagnosis part and a severity estimation part arranged hierarchically. It utilizes the latent features to exploit fault-related representations in the fault diagnosis task, improving the accuracy of fault severity estimation. FI-HCNN is developed based on stator current signals, which are commonly acquired for control purposes, making it easy to apply in practical settings. Experimental studies on mechanical motor faults, including eccentricity, have shown that FI-HCNN outperforms conventional methods and other hierarchical deep learning methods in fault severity estimation.
What are the advantages of using artificial neural networks for fault detection and diagnosis?5 answersArtificial neural networks offer several advantages for fault detection and diagnosis. They have the ability to handle complex systems with nonlinearity and large scale, making them suitable for diagnosing faults in distribution networks. Additionally, artificial neural networks exhibit robustness, fault tolerance, and strong self-learning ability, which are crucial for accurate fault detection and diagnosis in chemical processes. Furthermore, artificial neural networks can accurately detect and classify faults in DC distribution networks, even in the presence of noise and communication delay. Moreover, artificial neural networks, such as Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM), have been proven effective for fault detection and diagnosis in complex technical systems, such as pneumatic robots. Overall, the use of artificial neural networks enables early and accurate fault detection and diagnosis, improving the efficiency and safety of various systems.
What are the steps of artificial neural networks for fault detection and diagnosis?5 answersArtificial neural networks (ANNs) are used for fault detection and diagnosis in various systems. The steps involved in ANNs for fault detection and diagnosis are as follows. First, features are extracted from the system data using techniques like principal component analysis (PCA). Next, the ANN model is updated using the low-dimensional data obtained from PCA. The ANN is trained to identify different types of faults based on the updated model. The fault detection and diagnosis process involves two stages: model output estimation error stage and fault detection and diagnosis stage. In the model output estimation error stage, a nonlinear autoregressive exogenous neural network is used to model the system. In the fault detection and diagnosis stage, a standard feed-forward neural network is used to classify the type of fault based on the model output error estimates. The effectiveness of this approach has been demonstrated in various systems, including solar power plantsand pneumatic robots.
Training of artificial neural networks for fault detection and diagnosis in dynamic system?3 answersArtificial neural networks have been used for fault detection and diagnosis in dynamic systems. One approach is to use Long Short-Term Memory (LSTM) models, which are a type of recurrent neural network, to capture the dynamic characteristics of time series data. Another approach is to utilize feature reduction techniques such as Principal Component Analysis (PCA) and Non-Negative Matrix Factorization (NMF) as preprocessing steps in the machine learning pipeline. These techniques help mitigate the overfitting problem and improve the accuracy of fault detection. Additionally, the use of artificial neural networks in fault detection and diagnosis can lead to a reduction in production costs caused by malfunctioning equipment. Experimental results have shown that these methods outperform traditional approaches and can accurately identify different types of faults in dynamic systems.
- What methods and metrics are typically used to assess the performance of ANN-based fault recognition systems?5 answersArtificial Neural Network (ANN) based fault recognition systems are typically assessed using performance metrics such as mean square error, root mean square error, mean absolute error, linear regression, accuracy, and error rate. These metrics are used to evaluate the efficiency and accuracy of the ANN models for fault detection and classification. Additionally, graphical tools like regression error characteristic curves and sliding occurrence error curves are applied to verify the performance of the ANN architectures in terms of accuracy and promptness. The ANN models are also evaluated based on their adaptability to the problem at hand, training techniques such as backpropagation, and optimization techniques like Particle Swarm Optimization (PSO) for improving prediction accuracy.
Why is CNN and ANN preferred in detecting eccentricity fault in electrical machines?5 answersCNN and ANN are preferred in detecting eccentricity fault in electrical machines because they offer effective fault recognition and classification. These techniques utilize advanced signal processing algorithms to analyze the stator current and extract useful information for fault detection and characterization. Additionally, ANN-GA (Artificial Neural Networks - Genetic Algorithm) combined with ESPRIT method variants have been proposed for efficient fault recognition in real-time. The ESPRIT TLS variant allows for accurate discrimination of bearing faults even in the presence of noise, making it suitable for dataset preparation and ANN training. Furthermore, topological data analysis (TDA) methods have been developed for motor current signature analysis (MCSA) and applied to eccentricity fault detection. TDA can extract topological features from time-domain data and differentiate signals from different eccentricity levels, enabling the use of a simple regression model for fault level prediction.