A power quality detection and classification algorithm based on FDST and hyper-parameter tuned light-GBM using memetic firefly algorithm

Panigrahi, Rasmi Ranjan; Mishra, Manohar; Nayak, Janmenjoy; Shanmuganathan, Vimal; Naik, Bighnaraj; Jung, Youngae

doi:10.1016/j.measurement.2021.110260

Cited by 16 publications

(2 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As feature selection is a crucial step for classifier performance, it is needed to be taken care of. Unlike other classifiers LGBM [42][43][44] automatically take care of feature selection applying its internal GOSS and EFB mechanism, therefore used as base classifier of the proposed study. Similarly, detection time is also a much crucial performance parameter especially when real-time monitoring is performed and thus calculated in the proposed work.…”

Section: Comparative Studymentioning

confidence: 99%

“…Similarly, detection time is also a much crucial performance parameter especially when real-time monitoring is performed and thus calculated in the proposed work. In addition to that, [7] (Memetic Fire Fly Algorithm-MFFA), [42] (Improved Gray Wolf Optimization-IGWO) studies undergone the optimization of classifier parameters but suggested classifier is not validated in real time. Similarly, [40] has tested its PQ detection technique with data extracted from PQSCADA system but no in real-time mode.…”

Section: Comparative Studymentioning

confidence: 99%

See 1 more Smart Citation

Real Time Intelligent Detection of PQ Disturbances With Variational Mode Energy Features and Hybrid Optimized Light GBM Classifier

Mishra,

Mallick,

Gadanayak

et al. 2024

IEEE Access

View full text Add to dashboard Cite

The modern era power system is constantly undergoing constructive changes and implementations both in source and load side. Certainly, the distributed generators, unconventional/nonlinear loads, charging stations etc are mostly integrated through power electronics interfaces. As a result, frequent power quality disturbances appear in the system that is to be mitigated at the earliest. Since detection is the prerequisite for mitigation, therefore the article presents a novel intelligent power quality detection scheme to detect and classify the PQ Events. At first, the energy feature of the 5 band limited modes are calculated from variational mode decomposed voltage signals. Then the mode energy features are utilized to train a novel Hybrid Arithmetic Whale Optimized light gradient boosting machine classifier. A total of 15 different PQ events have been investigated and exceptional classification results have obtained with optimum computational complexity, both under noiseless and noisy conditions. Moreover, the accuracy of the proposed PQ classification schemes found to be towering against other related pre-published works. Finally, the ability of the proposed detection scheme is validated in real time though OPAL-RT 4510 and grid simulator hardware in loop setup.

show abstract

Section: Comparative Studymentioning

confidence: 99%