Article
Engineering, Electrical & Electronic
Utkarsh Kumar, Sukumar Mishra, Kalyan Dash
Summary: This article proposes a low-cost and less data-intensive methodology for detecting, localizing, and classifying faults in solar photovoltaic (SPV) systems. It utilizes a sensorless electronic circuit, an Internet of things (IoT)-based application, and a deep autoencoder-based semi-supervised learning module, followed by a hybrid support vector machine and logistic regression for fault classification. The methodology is validated in a laboratory-scale real-time setup of a grid-connected SPV system.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2023)
Review
Green & Sustainable Science & Technology
B. Li, C. Delpha, D. Diallo, A. Migan-Dubois
Summary: This review systematically studies the application of Artificial Neural Network (ANN) and hybridized ANN models for PV fault detection and diagnosis, extracting and analyzing the targeted PV faults, detectable faults, data types and amounts, model configurations, and FDD performance for each application. The main trends, challenges, and prospects for the application of ANN for PV FDD are identified and presented.
RENEWABLE & SUSTAINABLE ENERGY REVIEWS
(2021)
Article
Automation & Control Systems
Mansour Hajji, Mohamed-Faouzi Harkat, Abdelmalek Kouadri, Kamaleldin Abodayeh, Majdi Mansouri, Hazem Nounou, Mohamed Nounou
Summary: This paper aims to develop an improved FDD technique for PV system faults, utilizing principal component analysis for feature extraction and selection, and supervised machine learning classifiers for fault diagnosis. Experimental results confirm the feasibility and effectiveness of the proposed approaches for fault detection and diagnosis.
EUROPEAN JOURNAL OF CONTROL
(2021)
Article
Energy & Fuels
Mojgan Hojabri, Samuel Kellerhals, Govinda Upadhyay, Benjamin Bowler
Summary: This paper discusses the impact of faults in individual modules on the efficiency and reliability of the entire photovoltaic system. By using low-cost sensors and machine learning models, it is possible to achieve better fault detection performance. The results confirm the high classification accuracy of the neural network model in detecting faults.
Article
Computer Science, Artificial Intelligence
Jose L. Salazar Gonzalez, Juan A. Alvarez-Garcia, Fernando J. Rendon-Segador, Fabio Carrara
Summary: This study presents a semi-supervised learning approach based on conditioned cooperative student-teacher training, which utilizes Closed Circuit Television (CCTV) and weapon detection models to reduce violent assaults and homicides. The effectiveness of the approach is demonstrated by collecting a new firearms image dataset and comparing it with various learning techniques.
Article
Construction & Building Technology
Cheng Fan, Xuyuan Liu, Peng Xue, Jiayuan Wang
Summary: This study proposes a novel semi-supervised FDD method using neural networks, which adopts the self-training strategy for semi-supervised learning and has been tested for fault diagnosis and unseen fault detection. Statistical characterization of key learning parameters has been conducted through data experiments, showing that the method can effectively enhance model generalization performance by utilizing large amounts of unlabeled data.
ENERGY AND BUILDINGS
(2021)
Review
Computer Science, Artificial Intelligence
Jose Marcio Duarte, Lilian Berton
Summary: A large amount of data is generated daily, leading to challenges in handling big data. One of the challenges is in text mining, particularly text classification. Semi-supervised learning (SSL), which utilizes labeled and unlabeled data, has become increasingly important in this field. This paper aims to fill the gap by providing an up-to-date review of SSL for text classification, analyzing the application domain, datasets, languages, text representations, machine learning algorithms, evaluation metrics, and future trends.
ARTIFICIAL INTELLIGENCE REVIEW
(2023)
Article
Computer Science, Artificial Intelligence
Zijia Zhang, Yaoming Cai, Wenyin Gong
Summary: This paper presents a novel semi-supervised learning framework, Graph Convolutional Extreme Learning Machines (GCELM), for handling graph data in non-Euclidean domains. The proposed methods achieve significantly better results than previous methods on 36 benchmark datasets, thanks to the use of random graph convolution and a voting ensemble strategy.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Computer Science, Artificial Intelligence
Zhenpeng Lao, Deqiang He, Zhenzhen Jin, Chang Liu, Hui Shang, Yiling He
Summary: The turnout switch machine is crucial for train safety in the signal system. However, the lack of labeled fault data in real scenes hampers the diagnostic precision and generalization performance of the fault identification model. To address this, a semi-supervised weighted prototypical network (SSWPN) is proposed. It utilizes a dual-scale neural network (DSNN) and a semi-supervised weighted prototype updating strategy to enhance fault diagnosis. Experimental results demonstrate the robustness and generalization of SSWPN, outperforming eight other methods in data scarcity scenarios and providing a theoretical basis for few-shot fault diagnosis of the switch machine.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Energy & Fuels
Paria Movahed, Saman Taheri, Ali Razban
Summary: Long-term operation of HVAC systems can lead to failures, higher energy consumption, and maintenance costs. Fault detection diagnostic (FDD) is commonly used to prevent malfunctions, and machine learning methods have gained interest due to their high accuracy. However, existing studies suffer from biased classification algorithms and high false positives. To address these challenges and improve diagnostic performance, this study proposes a novel data-driven framework using principal component analysis, time series anomaly detection, and random forest.
Article
Energy & Fuels
Jiong Yang, Fanyong Cheng, Zhi Liu, Maxwell Mensah Duodu, Mingyan Zhang
Summary: This paper proposes a data-driven model based on kernel principal component analysis (KPCA) to achieve accurate and early detection as well as economical isolation of faults in vehicle battery systems. The model is trained using Bayesian Optimization iterations and a large amount of unlabeled data to overcome the challenge of hyperparameter selection. The proposed method also adopts a unified contribution graph based on partial differentiation of KPCA to build a reasonable fault isolation scheme.
Article
Automation & Control Systems
Yong Feng, Jinglong Chen, Tianci Zhang, Shuilong He, Enyong Xu, Zitong Zhou
Summary: In this paper, a semi-supervised meta-learning network with attention mechanism is proposed for few-shot fault diagnosis in mechanical systems. The method utilizes unlabeled data to improve fault recognition and achieves outstanding adaptability in different situations, as demonstrated through experiments with bearing vibration datasets.
Article
Engineering, Mechanical
Xinya Wu, Yan Zhang, Changming Cheng, Zhike Peng
Summary: Accurate fault diagnosis is crucial for the safe operation of rotating machinery. The proposed method, hybrid classification autoencoder, utilizes both labeled and unlabeled data for training and achieved high diagnostic accuracies in experiments with minimal labeled data. This approach shows potential for more efficient fault diagnosis in the future.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2021)
Article
Computer Science, Information Systems
Mona Nashaat, Aindrila Ghosh, James Miller, Shaikh Quader
Summary: This article presents a classification model that applies ensemble learning and data-driven rectification to deal with inaccurate and incomplete supervised datasets. The model detects noisy data points and improves the performance of the final classifier through ensemble learning and meta-learning methods.
ACM TRANSACTIONS ON KNOWLEDGE DISCOVERY FROM DATA
(2022)
Article
Environmental Sciences
Ehtisham Lodhi, Fei-Yue Wang, Gang Xiong, Lingjian Zhu, Tariku Sinshaw Tamir, Waheed Ur Rehman, M. Adil Khan
Summary: This research article presents a novel deep stack-based ensemble learning (DSEL) approach for diagnosing PV array faults. The DSEL approach combines deep neural network, long short-term memory, and Bi-directional long short-term memory models as base learners, with multinomial logistic regression as a meta-learner. The findings show that the proposed approach outperforms other techniques, achieving high accuracy for fault detection in both noiseless and noisy data.