Cascade ensemble learning for multi-level reliability evaluation

Aerospace Science and Technology, Journal Year: 2024, Volume and Issue: 148, P. 109101 - 109101

Published: April 1, 2024

Language: Английский

---

An Interpretable XGBoost-SHAP Machine Learning Model for Reliable Prediction of Mechanical Properties in Waste Foundry Sand-Based Eco-Friendly Concrete

Results in Engineering, Journal Year: 2025, Volume and Issue: unknown, P. 104307 - 104307

Published: Feb. 1, 2025

Language: Английский

---

Solar Photovoltaic System Performance Improvement Using a New Fault Identification Technique

Electric Power Components and Systems, Journal Year: 2023, Volume and Issue: 52(1), P. 42 - 54

Published: July 24, 2023

Identifying faults in the photovoltaic (PV) arrays is very much essential improving PV system's safety and reliability. Solar operate with non-linear characteristics, installed maximum power point trackers (MPPT's), blocking diodes cause mismatch levels. Line-to-line line-to-ground are identified, faulted circuits isolated by means of over current protection devices (OCPD) ground fault (GFPD). In order to improve accuracy detection, artificial intelligence (AI)-based techniques like Fuzzy inference, wavelet, support vector machine, k-nearest neighbors used. The drawback AI-based (1) requirement large dataset for effective identification also show incompatibility if there low irradiation (2) require a larger number voltage sensors. An experimental setup 160 W, 4 × solar array having modules (SPB) subjected different conditions (CS), identified using minimum that not detected conventional methods this proposed method, gain due around 152% which 97 W array.

Language: Английский

---

Artificial Intelligence in Photovoltaic Fault Identification and Diagnosis: A Systematic Review

Energies, Journal Year: 2023, Volume and Issue: 16(21), P. 7417 - 7417

Published: Nov. 3, 2023

Photovoltaic (PV) fault detection is crucial because undetected PV faults can lead to significant energy losses, with some cases experiencing losses of up 10%. The efficiency systems depends upon the reliable and diagnosis faults. integration Artificial Intelligence (AI) techniques has been a growing trend in addressing these issues. goal this systematic review offer comprehensive overview recent advancements AI-based methodologies for detection, consolidating key findings from 31 research papers. An initial pool 142 papers were identified, which selected in-depth following PRISMA guidelines. title, objective, methods, each paper analyzed, focus on machine learning (ML) deep (DL) approaches. ML DL are particularly suitable their capacity process analyze large amounts data identify complex patterns anomalies. This study identified several AI used systems, ranging classical methods like k-nearest neighbor (KNN) random forest more advanced models such as Convolutional Neural Networks (CNNs). Quantum circuits infrared imagery also explored potential solutions. analysis found that models, general, outperformed traditional accuracy efficiency. shows have evolved increasingly applied detection. offers high effectiveness. After reviewing studies, we proposed an Network (ANN)-based method classification.

Language: Английский

---

Enhancing microgrid forecasting accuracy with SAQ-MTCLSTM: A self-adjusting quantized multi-task ConvLSTM for optimized solar power and load demand predictions

Energy Conversion and Management X, Journal Year: 2024, Volume and Issue: unknown, P. 100767 - 100767

Published: Oct. 1, 2024

Language: Английский

---

A Novel Deep Stack-Based Ensemble Learning Approach for Fault Detection and Classification in Photovoltaic Arrays

Remote Sensing, Journal Year: 2023, Volume and Issue: 15(5), P. 1277 - 1277

Published: Feb. 25, 2023

The widespread adoption of green energy resources worldwide, such as photovoltaic (PV) systems to generate and renewable power, has prompted safety reliability concerns. One these concerns is fault diagnostics, which needed manage the output PV systems. Severe faults make detecting challenging because drastic weather circumstances. This research article presents a novel deep stack-based ensemble learning (DSEL) approach for diagnosing array faults. DSEL compromises three deep-learning models, namely, neural network, long short-term memory, Bi-directional base learners To better analyze arrays, we use multinomial logistic regression meta-learner combine predictions learners. study considers open circuits, short partial shading, bridge, degradation faults, incorporation MPPT algorithm. algorithm offers reliable, precise, accurate PV-fault diagnostics noiseless noisy data. proposed quantitatively examined compared eight prior machine-learning deep-learning-based classification methodologies by using simulated dataset. findings show that outperforms other techniques, achieving 98.62% accuracy detection with data 94.87% revealed retains strong generalization potential while enhancing prediction accuracy. Hence, detects categorizes more efficiently, reliably, accurately.

Language: Английский

---

MLPNN and Ensemble Learning Algorithm for Transmission Line Fault Classification

International Transactions on Electrical Energy Systems, Journal Year: 2025, Volume and Issue: 2025(1)

Published: Jan. 1, 2025

Recently, Bangladesh experienced a system loss of 11.11%, leading to significant power cuts, largely due faults in transmission lines. This paper proposes the XGBoost machine learning method for classifying electric line faults. The study compares multiple approaches, including ensemble methods (decision tree, random forest, XGBoost, CatBoost, and LightGBM) multilayer perceptron neural network (MLPNN), under various conditions. is modeled using Simulink algorithms. In IEEE 3‐bus system, all types achieve approximately 99% accuracy imbalanced noisy data states, respectively, except CatBoost decision classification line, ground, ground faults, no fault. However, although gain accuracy, assessing performance results indicates that model most effective fault among tested, as it showed best state’s contributing development more reliable efficient detection methodologies networks.

Language: Английский

---

Towards Accurate and Reliable Fault Diagnosis in PV Systems: Techniques, Challenges, and Future Directions

Process Safety and Environmental Protection, Journal Year: 2025, Volume and Issue: unknown, P. 107217 - 107217

Published: April 1, 2025

Language: Английский

---

Forecasting capacitor banks for improving efficiency of grid-integrated PV plants: A machine learning approach

Energy Reports, Journal Year: 2024, Volume and Issue: 13, P. 140 - 160

Published: Dec. 9, 2024

Language: Английский

---

Using SegFormer for Effective Semantic Cell Segmentation for Fault Detection in Photovoltaic Arrays

IEEE Journal of Photovoltaics, Journal Year: 2024, Volume and Issue: 15(2), P. 320 - 331

Published: Sept. 5, 2024

Photovoltaic (PV) industries are susceptible to manufacturing defects within their solar cells. To accurately evaluate the efficacy of PV modules, identification is imperative. Conventional industrial defect inspections predominantly rely on highly skilled inspectors conducting manual assessments, leading sporadic and subjective outcomes. Deep-learning-based fault detection in or cells has emerged as a primary research area due its superior efficiency applicability. Hence, this study introduces SegFormer-based framework automate visual inspection process complete with pseudocolorization. The proposed effectively classifies into five categories: crack defects, front grid interconnect contact corrosion bright disconnect. Moreover, comparative analysis performed between SegFormer model state-of-the-art algorithms, such Deeplab v3, UNET, v3+, PAN, PSPNet, feature pyramid network (FPN). experimental results reveal that achieves encouraging performance, pixelwise accuracy 96.24%, weighted F1-score 96.22%, an unweighted 81.96%, mean intersection over union 56.54%, outperforming other existing methods.

Language: Английский

---