Photovoltaic Fault Research Articles

Intelligence in Photovoltaic Fault Identification and Diagnosis: A Systematic Review

The increasing global demand for renewable energy has propelled the adoption of photovoltaic systems as a key component of sustainable energy infrastructure. Undetected photovoltaic system faults can lead to significant energy losses, often exceeding 10%, necessitating efficient fault detection and diagnosis. Artificial intelligence, particularly machine learning and deep learning, offers promising solutions for real-time, high-volume fault detection and complex pattern recognition in PV systems. This research analyzes various PV fault detection studies, examining their objectives, methods, results, and the prevalence of ML/DL approaches. The analysis highlights the application of both classical ML algorithms, such as K-Nearest Neighbors and Random Forest, and advanced DL models, including Convolutional Neural Networks, for PV fault diagnosis.

Intelligence in Photovoltaic Fault Identification and Diagnosis: A Systematic Review

Undetected photovoltaic system faults can lead to significant energy losses, often exceeding 10%, necessitating efficient fault detection and diagnosis. Artificial intelligence, particularly machine learning and deep learning, offers promising solutions for real-time, high-volume fault detection and complex pattern recognition in PV systems. This research analyzes various PV fault detection studies, examining their objectives, methods, results, and the prevalence of ML/DL approaches. The analysis highlights the application of both classical ML algorithms, such as K-Nearest Neighbors and Random Forest, and advanced DL models, including Convolutional Neural Networks, for PV fault diagnosis.

Comparative study of real-time photovoltaic fault diagnosis using artificial intelligence: Fuzzy logic and neural network approaches

ABSTRACT Solar photovoltaic (PV) systems are becoming increasingly popular for renewable energy production. However, due to environmental and operational conditions, various faults can occur in PV modules, which can cause a significant reduction in system performance. This study suggests employing two distinct approaches, Fuzzy Logic (FL) and Artificial Neural Networks (ANN), to diagnose defects in photovoltaic (PV) systems. Both the simulation and measuring are used to compare the performance of various strategies. The simulation was performed using MATLAB/Simulink, while the experiment was carried out using the dSPACE DS1104 platform to apply the diagnostic model built in the Matlab/Simulink® program. The suggested approaches have been verified using an experimental database of climatic and electrical attributes obtained from a photovoltaic (PV) panel situated at the LGEB Laboratory of the University of Biskra in Algeria. For six single- and multi-fault types, partial shading, soiling, SC of one by-pass diode, SC of two by-pass diodes, by-pass diode shunted and shading & by-pass diode disconnected. A dSPACE DS1104 platform which shows the results and informs the user about the state of the PV panel has also been developed. The results of the simulation show that both FL and ANN methods can accurately detect and diagnose the six types of faults by 100% and 99.7%, respectively. The study demonstrates the effectiveness of both FL and ANN methods for PV fault diagnosis. However, in the experimental tests, the ANN method shows better performance in terms of accuracy compared to FL by 99.6% and 99.2%, respectively, and speed takes only 1.04s, while FL consumes 3.02s and could be a more practical solution for real-time surveillance PV fault diagnosis

Anomaly detection of photovoltaic power generation based on quantile regression recurrent neural network

Distributed photovoltaic (PV) power generation systems are widely spread. Moreover, due to the randomness of meteorological conditions and the complexity of installation environments, it is difficult to eliminate the interference of factors such as meteorological fluctuations in the monitoring of abnormal states of PV equipment. Based on this, this paper proposes a PV power generation anomaly detection method based on Quantile Regression Recurrent Neural Network (QRRNN). First, the characteristics of solar irradiance on clear days are analyzed, and the clear day masking method is used to eliminate the interference of cloudy and rainy weather. Then, the output correlation of different power stations is analyzed to obtain PV stations with high output correlation as the horizontal reference, which is used to exclude interferences such as permanent faults at the power stations. At the same time, vertical comparison of the output curves of the station under test on different clear days is conducted to eliminate interference factors such as weather and environmental conditions. Subsequently, the metered active power output data, which is free from interference, is input into the QRRNN model to obtain the normal active power output range of the PV. The power threshold of the normal output range is utilized to identify anomalies in PV power generation. Finally, simulation analysis of actual PV system data is conducted, and the results show that the method can effectively identify PV power generation anomalies and has high accuracy in PV fault detection.

RAM-YOLOv8: A multi-fault detection method for photovoltaic module

ABSTRACT Given the constraints in fault type identification, localization precision, and detection efficiency inherent in current photovoltaic fault detection methodologies, this paper introduces an innovative detection approach integrating RAM-YOLOv8 with the Swin Transformer. To mitigate the limitations in model extraction capabilities during the detection process and circumvent the vanishing gradient issue encountered during backward information propagation, this study introduces a residual aggregation module as a substitute for the original C2F module, thereby augmenting feature extraction efficiency. Furthermore, within the feature fusion network, the Swin attention mechanism fusion module is designed by integrating the Swin Transformer’s architectural concept to minimize the loss of weak defect information and suppress irrelevant background interference. The model achieves a detection accuracy of 88.5% on the dataset, representing a 7% improvement over the original YOLOv8 algorithm, with a detection rate of 83.33 FPS on an RTX2080 GPU. This method offers a robust and efficient fault detection solution for PV systems, rendering it a practical approach for enhancing system performance and reducing maintenance costs.

A Novel Strategy for Multitype Fault Diagnosis in Photovoltaic Systems Using Multiple Regression Analysis and Support Vector Machines

This study focuses on analyzing common fault types in photovoltaic (PV) modules, employing fault diagnosis methods based on machine learning technology to enhance the accuracy and efficiency of diagnosing faults in solar power systems. Initially, we collected relevant data from the solar power system and used data analysis techniques to identify system faults, designing a human-machine monitoring interface for practical application. Furthermore, the experimental results proved that the system could accurately identify eight major types of faults, including solar panel output circuits, energy storage batteries, maximum power point tracking (MPPT) controllers, inverters, dust accumulation, loosening of mounting rack screws, damage to the mounting rack foundation, and deformation of the mounting rack structure. Particularly in the detection of dust accumulation, we developed a new method of estimating power generation from multiple regression analysis (MRA), which closely aligns the estimated power output with the actual power output, highlighting the significant impact of dust accumulation on the efficiency of solar power systems. Next, by integrating voltmeters and support vector machines (SVM) into the solar PV array modules, we are able to quickly and accurately measure and locate short-circuit and open-circuit faults in bypass diodes. Ultimately, the proposed PV fault diagnosis strategy includes diagnostics for dust accumulation and mounting frame faults, making it particularly suitable for areas with severe air pollution and frequent earthquakes, providing a comprehensive fault diagnosis solution.

An ensemble learning framework for snail trail fault detection and diagnosis in photovoltaic modules

This research proposes a method for detecting subtle faults named snail trails for their visual similarity with the trail of a snail in photovoltaic modules. Snail trails do not significantly reduce panel performance but they are the main cause of serious panel deterioration such as microcracks and delamination and can go so far as to set the panel on fire. To detect these faults, this research uses an ensemble learning framework, named ensemble learning for diagnosis, which combines several complementary learning algorithms, namely Support Vector Machines, K-Nearest Neighbors, and Decision Trees. A set of features is obtained by extracting the time–frequency characteristics and statistics from the photovoltaic current signal of the photovoltaic panel. This is followed by a feature selection and dimensionality reduction step that delivers the input to the learning algorithms. The approach presented in this study is experimentally validated, independently for the 4 seasons of the year, with data from a real photovoltaic string of 16 panels. The results demonstrate that the proposed approach can efficiently classify healthy panels and panels with snail trails efficiently. Interestingly, the method only requires the electrical current signal, measured on panels with data acquisition systems that are standard in the photovoltaic industry. The genericity of the approach makes it a good candidate for detecting other photovoltaic faults and for solving diagnosis problems in other domains.

PVF-10: A high-resolution unmanned aerial vehicle thermal infrared image dataset for fine-grained photovoltaic fault classification

Accurate identification of faulty photovoltaic (PV) modules is crucial for the effective operation and maintenance of PV systems. Deep learning (DL) algorithms exhibit promising potential for classifying PV fault (PVF) from thermal infrared (TIR) images captured by unmanned aerial vehicle (UAV), contingent upon the availability of extensive and high-quality labeled data. However, existing TIR PVF datasets are limited by low image resolution and incomplete coverage of fault types. This study proposes a high-resolution TIR PVF dataset with 10 classes, named PVF-10, comprising 5579 cropped images of PV panels collected from 8 PV power plants. These classes are further categorized into two groups according to the repairability of PVF, with 5 repairable and 5 irreparable classes each. Additionally, the circuit mechanisms underlying the TIR image features of typical PVF types are analyzed, supported by high-resolution images, thereby providing comprehensive information for PV operators. Finally, five state-of-the-art DL algorithms are trained and validated based on the PVF-10 dataset using three levels of resampling strategy. The results show that the overall accuracy (OA) of these algorithms exceeds 83%, with the highest OA reaching 93.32%. Moreover, the preprocessing procedure involving resampling and padding strategies are beneficial for improving PVF classification accuracy using PVF-10 datasets. The developed PVF-10 dataset is expected to stimulate further research and innovation in PVF classification.

Detection and Classification of Physical and Electrical Fault in PV Array System by Random Forest-Based Approach

The importance of solar photovoltaic (PV) systems has increased over the past ten years due to the solar PV industry's explosive growth. To ensure the reliable, safe, and efficient operation of residential PV systems, fault detection is crucial. Early classification of faults can improve PV system performance and reduce damage and energy loss. Many recent studies have focused on classifying and detecting PV faults but most of them are limited to specific reasons like Real-world data can be restricted, unbalanced, or include noise, all of which may decrease the effectiveness of ML models. This paper proposes a method for identifying and classifying both physical and electrical faults in the PV array system applying a machine learning (Random Forest) model to that is trained using a synthetic photovoltaic training database. Make use of a synthetic PV database opens the door to a more precise, effective, and scalable PV system by addressing the limitations of real-world data. MATLAB is used to create a synthetic database while scikit-learn tool in Jupyter Notebook is used to train an ML model are the two main steps in this paper. The performance of the proposed model is compared with the existing ML model and achieves the most effective algorithm offering higher accuracy in detection of 98.6% and classification accuracy is 94.2% for both physical and electrical faults after being successfully tested on real-world datasets and trained on historical data from the PV array system (PV Database).

In-Depth Review of YOLOv1 to YOLOv10 Variants for Enhanced Photovoltaic Defect Detection

This review presents an investigation into the incremental advancements in the YOLO (You Only Look Once) architecture and its derivatives, with a specific focus on their pivotal contributions to improving quality inspection within the photovoltaic (PV) domain. YOLO’s single-stage approach to object detection has made it a preferred option due to its efficiency. The review unearths key drivers of success in each variant, from path aggregation networks to generalised efficient layer aggregation architectures and programmable gradient information, presented in the latest variant, YOLOv10, released in May 2024. Looking ahead, the review predicts a significant trend in future research, indicating a shift toward refining YOLO variants to tackle a wider array of PV fault scenarios. While current discussions mainly centre on micro-crack detection, there is an acknowledged opportunity for expansion. Researchers are expected to delve deeper into attention mechanisms within the YOLO architecture, recognising their potential to greatly enhance detection capabilities, particularly for subtle and intricate faults.

Leveraging dynamic power benchmarks and CUSUM charts for enhanced fault detection in distributed PV systems

Faults in photovoltaic (PV) systems are common during their operational lifetime. Existing PV fault detection methods, which are primarily designed for large-scale PV fields, struggle with distributed systems due to their reliance on on-site weather sensors and inability to handle complex local shading effects on PV performance in the built environment. This paper presents a fault detection scheme specifically applicable to distributed PV systems that solely relies on monitored AC output and remote weather sources. Without requiring additional equipment or detailed information on the PV configuration and local shading conditions, the proposed method consists of two steps. First, historical monitoring data is used in a bootstrap fashion to develop machine learning models that establish baselines for normal PV output and corresponding estimation uncertainty. Then, a dynamic PV power benchmark constructed based on the 3-sigma rule, and cumulative sum (CUSUM) control charts are employed simultaneously to detect system malfunctions. Through experimental verification on actual distributed PV systems, the effectiveness of the proposed method is assessed for four categories of frequently observed malfunctions. The comprehensive experimental results indicate the considerable efficiency of the proposed sensorless method in detecting both serious PV malfunctions and minor faults with a severity as small as 4.2%.

Intrinsic performance loss rate: Decoupling reversible and irreversible losses for an improved assessment of photovoltaic system performance

AbstractSolar electricity is set to play a pivotal role in future energy systems. In view of a market that may soon reach the terawatt (TW) scale, a careful assessment of the performance of photovoltaic (PV) systems becomes critical. Research on PV fault detection and diagnosis (FDD) focuses on the automated identification of faults within PV systems through production data, and long‐term performance evaluations aim to determine the performance loss rate (PLR). However, these two approaches are often handled separately, resulting in a notable gap in the field of reliability. Within PV system faults, one can distinguish between permanent, irreversible effects (e.g. bypass diode breakage, delamination and cell cracks) and transient, reversible losses (e.g. shading, snow and soiling). Reversible faults can significantly impact (and bias) PLR estimates, leading to wrong judgements about system or component performance and misallocation of responsibilities in legal claims. In this work, the PLR is evaluated by applying a fault detection procedure that allows the filtering of shading, snow and downtime. Compared with standard filtering methods, the addition of an integrated FDD analysis within PLR pipelines offers a solution to avoid the influence of reversible effects, enabling the determination of what we call the intrinsic PLR (i‐PLR). Applying this method to a fleet of PV systems in the built environment reveals four main PLR bias scenarios resulting from shading losses. For instance, a system with increasing shading over time exhibits a PLR of −1.7%/year, which is reduced to −0.3%/year when reversible losses are filtered out.

Photovoltaic fault detection based on infrared and visible image augmentation and fusion

Photovoltaic fault detection based on infrared and visible image augmentation and fusion

Worldwide scientific landscape on fires in photovoltaic

The rapid growth of photovoltaic (PV) technology in recent years called for a comprehensive assessment of the global scientific landscape on fires associated with PV energy installations. This study examines the scientific literature indexed in Scopus from 1983 to 2023. It reveals a striking increase in output since 2011, with nearly one hundred publications in the most recent year under review. This growth of interest has occurred in parallel with the global expansion of photovoltaics. The majority of studies in this field are classified as engineering, with 34% of publications in this area. The USA leads the way with over 160 publications, followed by China with 125. Two institutions in the USA are particularly prominent in this field: Sandia National Laboratories in New Mexico with 22 publications, and the National Renewable Energy Laboratory in Colorado with 16 publications. The second institution is the University of Science and Technology of China, which has published 17 articles on the subject. A close examination of the evolution of keywords reveals a remarkable transformation in the scientific landscape over the past 10 years, from 2013 to 2023. The evolution of keywords suggests a maturation in the understanding of fire risks associated with photovoltaic energy. A total of seven scientific communities have been identified in which these works are grouped according to their keywords. These include Fire and Energy Storage, PV faults, Fire resistance, Fire hazard, Fire detectors, Deep learning, and Fire safety. It has been found that fires caused by PV installations are not listed as a cause of fire starts. This should be taken into account when conducting preventive analyses of this potential danger, particularly in light of the possible development of agrivoltaics, where facilities will be mainly located in the natural environment.

PV Panel Model Parameter Estimation by Using Particle Swarm Optimization and Artificial Neural Network.

Photovoltaic (PV) panels are one of the popular green energy resources and PV panel parameter estimations are one of the popular research topics in PV panel technology. The PV panel parameters could be used for PV panel health monitoring and fault diagnosis. Recently, a PV panel parameters estimation method based in neural network and numerical current predictor methods has been developed. However, in order to further improve the estimation accuracies, a new approach of PV panel parameter estimation is proposed in this paper. The output current and voltage dynamic responses of a PV panel are measured, and the time series of the I-V vectors will be used as input to an artificial neural network (ANN)-based PV model parameter range classifier (MPRC). The MPRC is trained using an I-V dataset with large variations in PV model parameters. The results of MPRC are used to preset the initial particles' population for a particle swarm optimization (PSO) algorithm. The PSO algorithm is used to estimate the PV panel parameters and the results could be used for PV panel health monitoring and the derivation of maximum power point tracking (MMPT). Simulations results based on an experimental I-V dataset and an I-V dataset generated by simulation show that the proposed algorithms can achieve up to 3.5% accuracy and the speed of convergence was significantly improved as compared to a purely PSO approach.

Residual learning-based robotic image analysis model for low-voltage distributed photovoltaic fault identification and positioning.

With the fast development of large-scale Photovoltaic (PV) plants, the automatic PV fault identification and positioning have become an important task for the PV intelligent systems, aiming to guarantee the safety, reliability, and productivity of large-scale PV plants. In this paper, we propose a residual learning-based robotic (UAV) image analysis model for low-voltage distributed PV fault identification and positioning. In our target scenario, the unmanned aerial vehicles (UAVs) are deployed to acquire moving images of low-voltage distributed PV power plants. To get desired robustness and accuracy of PV image detection, we integrate residual learning with attention mechanism into the UAV image analysis model based on you only look once v4 (YOLOv4) network. Then, we design the sophisticated multi-scale spatial pyramid fusion and use it to optimize the YOLOv4 network for the nuanced task of fault localization within PV arrays, where the Complete-IOU loss is incorporated in the predictive modeling phase, significantly enhancing the accuracy and efficiency of fault detection. A series of experimental comparisons in terms of the accuracy of fault positioning are conducted, and the experimental results verify the feasibility and effectiveness of the proposed model in dealing with the safety and reliability maintenance of low-voltage distributed PV systems.

Intelligent fault diagnosis of photovoltaic systems based on deep digital twin

The energy loss and substantial costs associated with faults in photovoltaic (PV) systems impose significant limitations on their efficiency and reliability. Addressing current issues in PV fault diagnosis such as the lack of typical fault data, imbalanced data distribution, and poor diagnostic performance, this paper proposes an intelligent fault diagnosis method for PV systems, deep digital twins (DDT) with information gain stacking sparse autoencoders (IGSSAEs). Initially, the method designs a novel DDT modeling framework tailored to actual PV system specifications. This framework utilizes a mechanism simulation model to generate typical data under various states. Simultaneously, a deep data model is constructed to learn the distribution characteristics of the mechanism model and complete data diversification, achieving the fusion and complementation of data from both models. Subsequently, a diagnostic network using IGSSAE is introduced. This network utilizes information gain ratio to assess feature classification contributions, enabling automatic feature selection. Based on the input features, a stacked sparse autoencoder fault classification network is designed, incorporating multi-level feature compression to enhance the model’s stability and diagnostic accuracy. Finally, a case study is conducted using a 250 kW grid-connected PV system, thoroughly validating the method’s effectiveness with a diagnostic accuracy of 98.4%.

Recent advances in fault detection techniques for photovoltaic systems: An overview, classification and performance evaluation

Photovoltaic (PV) arrays are typically built outdoors in severe conditions and are vulnerable to a variety of defects, which will negatively impact their efficiency and reduce the system's performance as a whole. Therefore, implementing efficient fault identification and diagnosis is crucial to increase the productivity and efficiency of PV installations, ensure their safe operation and reduce maintenance costs. In this article, the types and causes of numerous faults that arise in PV systems are swiftly examined. Additionally, a number of the most recent methods suggested in the literature for PV fault diagnostics are reviewed and assessed. We only cover methods for locating faults on the DC side of the PV system. Based on the review that was conducted, suggestions for the direction of future studies are given. Finally, this document serves as a significant manual that offers relevant information on methods of detection and diagnosis for PV systems.

Implementation of Caputo type fractional derivative chain rule on back propagation algorithm

Fractional gradient computation is a challenging task for neural networks. In this study, the brief history of fractional derivation is abstracted, and the core framework of the Faà di Bruno formula is implemented to the fractional gradient computation problem. As an analytical approach to solve the chain rule problem of fractional derivatives, the use of the Faà di Bruno formula for the Caputo-type fractional derivative is proposed. When the fractional gradient is calculated using the proposed approach, the problem of determining the bounds of the formula for calculating the Caputo derivative is addressed. As a consequence, the fundamental problem with the fractional back-propagation algorithm is solved analytically, paving the way for the use of any differentiable and integrable activation function in case they are stable. The development of the algorithm and its practical implementation on the photovoltaic fault detection data-set is investigated. The reliability of the neural network metrics is also investigated using analysis of variance (ANOVA), the results are then presented to decide which are the best metrics and the best fractional order. Ordinary back-propagation, fractional back-propagation with and without L2 regularization and momentum are presented in the results to show the advantages of using L2 regularization and momentum in fractional order gradient. Consequently, the test metrics are reliable but cannot be separated from each other. By changing the batch size from 2 to full batches, the proposed system performs better with bigger batches, but adaptive moment estimation (ADAM) performs better with small batches. The cross validation shows the performance of back propagate ion neural networks have better performance than the ordinary neural networks. It is interesting that the best order for a specific data-set cannot be determined because it depends on the batch size, number of epochs and the cross-validation folds .

A Survey of Photovoltaic Panel Overlay and Fault Detection Methods

Photovoltaic (PV) panels are prone to experiencing various overlays and faults that can affect their performance and efficiency. The detection of photovoltaic panel overlays and faults is crucial for enhancing the performance and durability of photovoltaic power generation systems. It can minimize energy losses, increase system reliability and lifetime, and lower maintenance costs. Furthermore, it can contribute to the sustainable development of photovoltaic power generation systems, which can reduce our reliance on conventional energy sources and mitigate environmental pollution and greenhouse gas emissions in line with the goals of sustainable energy and environmental protection. In this paper, we provide a comprehensive survey of the existing detection techniques for PV panel overlays and faults from two main aspects. The first aspect is the detection of PV panel overlays, which are mainly caused by dust, snow, or shading. We classify the existing PV panel overlay detection methods into two categories, including image processing and deep learning methods, and analyze their advantages, disadvantages, and influencing factors. We also discuss some other methods for overlay detection that do not process images to detect PV panel overlays. The second aspect is the detection of PV panel faults, which are mainly caused by cracks, hot spots, or partial shading. We categorize existing PV panel fault detection methods into three categories, including electrical parameter detection methods, detection methods based on image processing, and detection methods based on data mining and artificial intelligence, and discusses their advantages and disadvantages.