Condition Monitoring for Power Converters via Deep One-Class Classification

Marković, Nikola; Vahle, Daniel; Staudt, V.; Kolossa, Dorothea

doi:10.1109/icmla52953.2021.00244

Cited by 3 publications

(2 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This information is then used to train a Machine Learning (ML) algorithm for RUL prediction of the entire converter. In [9] proposes a method to identify problems in power converters early, especially for modular multilevel converters (MMCs). Method involves training a special type of artificial intelligence (AI) called a one-class classifier.…”

Section: A Literature Reviewmentioning

confidence: 99%

A Machine Learning-based Solution for Monitoring of Converters in Smart Grid Application

Sadiq,

Mallek,

Rehman

et al. 2024

IJACSA

View full text Add to dashboard Cite

The integration of renewable energy sources and the advancement of smart grid technologies have revolutionized the power distribution landscape. As the smart grid evolves, the monitoring and control of power converters play a crucial role in ensuring the stability and efficiency of the overall system. This research paper introduced a converter monitoring system in photovoltaic systems, the main concern is to protect the electrical system from disastrous failures that occur when the system is in operating condition. The reliability of the converters is significantly influenced by the degradation of their passive components, which can be characterized in various ways. For instance, the aging of inductors and capacitors can be characterized by a decrease in their inductance and capacitance values. Identifying which component is undergoing degradation and assessing whether it is in a critical condition or not, is crucial for implementing cost-effective maintenance strategies. This paper explores a set of classification algorithms, leveraging machine learning, trained on data collected from a Zeta converter simulated in Matlab Simulink. the report presents observations on how each algorithm effectively predicts the component and its condition and Graphical Performance Comparison for different ML Techniques serves as a crucial endeavor in evaluating and understanding the effectiveness of various ML approaches. The goal is to provide a comprehensive overview of how these techniques fare concerning criteria such as accuracy, precision, recall, F1 score, and Specificity among others. Quadratic Support Vector Machine (SVM) yields superior results compared to other machine learning techniques employed in training our dataset.

show abstract

Section: A Literature Reviewmentioning

confidence: 99%

A Machine Learning-based Solution for Monitoring of Converters in Smart Grid Application

Sadiq,

Mallek,

Rehman

et al. 2024

IJACSA

View full text Add to dashboard Cite

show abstract

“…A data-driven method for detecting capacitance changes in MMCs is proposed in [28]. There, the SM voltage and the arm current signals are used to train a one-class-classier for detecting changes in capacitance, and T 2 statistics are used to improve the reliability of the system.…”

Section: ) Parameter Changesmentioning

confidence: 99%

Hybrid Condition Monitoring for Power Converters: Learning-Based Methods With Statistical Guarantees

et al. 2023

Self Cite

View full text Add to dashboard Cite

In this paper, a new learning-based method is proposed for the early detection of changes of parameters in power converters. It circumvents the pertinent shortcomings of previous model-based methods, such as their need for acquiring switching signals, dependence on the control method, or the need for isolating a part of the system during monitoring and thus interfering with the system performance or its start-up time. Additionally, a downside of the learning-based approaches, which is that their performance depends on the quality of the measurements, is addressed through constructing hybrid models that combine the benets of both lines of development. Our approaches are evaluated with several types of features based on wavelet decomposition and empirical mode decomposition. Using these, an ANN-based classier is trained for fault detection. For achieving the nal decision on the presence or absence of a fault state, we propose the use of sequential hypothesis testing. This hybrid approach yields a much higher reliability than instantaneous ANNbased classication, while allowing for a statistically sound cross-temporal information integration and providing a controllable error bound for the probability of misclassications. The proposed method was evaluated on two data-sets recorded from a buck converter and an arm of a modular multilevel converter. The results show that, for both systems, the proposed method is capable of reliably recognizing changes of the system parameters. The potential for practical applications is shown through an implementation on a low-power-consumption microcontroller. INDEX TERMS Fault diagnosis, power converters, DC/DC Converter, MMC, machine learning, signal processing, hybrid models I. INTRODUCTIONC ONDITION monitoring of power converters in energy systems is of great importance, considering the wide range of applications in which they play a critical role, and their long operational time. While there are many dierent power converter topologies, these broadly share the same principles of operation via active and passive elements. Hence, very similar condition monitoring methods can, with some modications, be applied for dierent power converters.The specic condition monitoring technique proposed in this paper is applied to two dierent power converter topologies. The rst is a buck converter, a widely employed DC-DC power converter [1]. The second system is an arm of a of modular multilevel converter (MMC). These converters have gained great interest in academia as well as industry shortly after they were initially proposed [2], due to their modular design, their

show abstract

A Novel Unsupervised Outlier Detection Algorithm Based on Mutual Information and Reduced Spectral Clustering

Huang,

Liu,

et al. 2023

Electronics

View full text Add to dashboard Cite

Outlier detection is an essential research field in data mining, especially in the areas of network security, credit card fraud detection, industrial flaw detection, etc. The existing outlier detection algorithms, which can be divided into supervised methods and unsupervised methods, suffer from the following problems: curse of dimensionality, lack of labeled data, and hyperparameter tuning. To address these issues, we present a novel unsupervised outlier detection algorithm based on mutual information and reduced spectral clustering, called MISC-OD (Mutual Information and reduced Spectral Clustering—Outlier Detection). MISC-OD first constructs a mutual information matrix between features, then, by applying reduced spectral clustering, divides the feature set into subsets, utilizing the LOF (Local Outlier Factor) for outlier detection within each subset and combining the outlier scores found within each subset. Finally, it outputs the outlier score. Our contributions are as follows: (1) we propose a novel outlier detection method called MISC-OD with high interpretability and scalability; (2) numerous experiments on 18 benchmark datasets demonstrate the superior performance of the MISC-OD algorithm compared with eight state-of-the-art baselines in terms of ROC (receiver operating characteristic) and AP (average precision).

show abstract

Condition Monitoring for Power Converters via Deep One-Class Classification

Cited by 3 publications

References 29 publications

A Machine Learning-based Solution for Monitoring of Converters in Smart Grid Application

A Machine Learning-based Solution for Monitoring of Converters in Smart Grid Application

Hybrid Condition Monitoring for Power Converters: Learning-Based Methods With Statistical Guarantees

A Novel Unsupervised Outlier Detection Algorithm Based on Mutual Information and Reduced Spectral Clustering

Contact Info

Product

Resources

About