Search

• An autonomous monitoring platform for the identification of visual failures on PV arrays. • Practical flights by multi-copters to collect a data-set of diverse aerial images of PV strings. • An automatic extractor was developed to extract modules from collected aerial images of PV strings. • Feature extraction and up-sampling to pixel level output through a developed encoder-decoder network. • Position conversion from image pixels to earth coordinates using a developed mapping method. • Performance evaluation of the designed network using real aerial images from the PV strings. • A decision-making algorithm for categorizing affected modules into groups of high, medium and low risk. • The evaluation of the effect of birds' drops on the electrical performance of PV modules. This study presents an autonomous fault detection method for a wide range of common failures and defects which are visually visible on PV modules. In this paper, we focus especially on detection of bird's drops as a very typical defect on the PV modules. As a crucial prerequisite, a data-set of aerial imageries of the PV strings affected by bird's drops were collected through several experimental flight by multi-copters in order to train an accurate fully convolutional deep network. These images are divided into three groups, namely, training, testing, and validation parts. For the purpose of bird's drops segmentation, an improved encoder-decoder architecture is employed. In this regard, a modified VGG16 model is used as a backbone for the encoder part. The encoder of the network has a very flexible architecture that can be modified and trained for any other visual failure detection. Later on, extracted feature maps of images are imported into a decoder network to map the low resolution features to full resolution ones for pixel-wise segmentation. In addition, an image object positioning algorithm is presented to find the exact position of detected failures in local coordinate system. In a post-processing step, the detected damages are prioritized based on various parameters such as severity of shading and extent of impact on the PV module's output current. For further validation, different affected PV modules were characterized according to the output patterns of the classification step in order to accurately evaluate the effect of birds' drops and consequent shading on the parameters of PV modules based on their severity and location. Finally, the training and testing results demonstrate that the proposed FCN network is able to predict precisely covered pixels by bird's drops on PV modules at pixel level with average accuracies of 98% and 93% for training and testing, respectively. [ABSTRACT FROM AUTHOR]

• aIRT equipment equipped with a RGB camera improves inspection. • Post-flight image analysis provides a faster aIRT inspection. • The most common problem detected were hot spots caused by soiling and vegetation. • The most common defect in the power plants were disconnected cell substrings. • The defect with higher impact on the PV plant performance are disconnected strings. The uptime of utility-scale solar photovoltaic (PV) power plants is of utmost importance for meeting contractual energy yields and expected capacity factors. Aerial Infrared Thermography (aIRT) is a non-destructive, no-downtime, fast and cost-effective method for monitoring large-scale PV power plants and assisting in fault detection. The use of aIRT techniques aims at increasing the quality and service life of PV plants especially in sunny and developing countries such as Brazil, where there is a shortfall of specialised workforce and the costs for a detailed supervisory system of utility-scale PV power plants are high. This paper presents an analysis of an aIRT flight campaign over four utility-scale PV plants in the northeast of Brazil. Two types of measurement equipment have been tested and compared, resulting in more stability and efficiency using a commercially available solution. This solution was also equipped with a RGB camera that accelerated the inspections, since it helped to differentiate defects from hot spots caused by soiling and vegetation over the modules, which were common. Different methods for fault detection were also tested and it was concluded that post-flight image analysis provides faster results. The most common faults that can happen in the early operation of PV power plants and how operators should address and prevent them were also discussed. The most common defect detected during the campaign were disconnected cell substrings. However, disconnected strings had the most impact on the power plants energy performance. [ABSTRACT FROM AUTHOR]

• Statistical features based on time and frequency domain have been extracted from the operating current and voltage of PV arrays. • A weighted ensemble learning algorithm based on a Genetic Algorithm (GA) has been proposed for faults diagnosis in PV arrays. • The Lasso penalty algorithm has been used to find an appropriate subset of features to improve performance accuracy and reduce the dataset required in the training process. Conventional protection devices may not be able to diagnose the faults in Photovoltaic (PV) systems due to the nonlinear behavior of PV characteristics, its dependency on the operating environment, and operation of Maximum Power Point Tracking (MPPT) algorithms. To date, numerous studies have been carried out to overcome this challenge through Artificial Intelligence (AI) techniques. However, most of the AI-based techniques require a large dataset and also suffer overfitting problem. In this study, we propose an intelligent and automatic fault diagnosis method using less dataset for the training process through feature extraction and selection algorithms, as well as using an ensemble learning algorithm to classify open-circuit (OC) and line-line (LL) faults in PV systems. For this purpose, the proposed model firstly extracts the key features of the operating current and voltage of the PV arrays. Secondly, the Lasso penalty as an embedded feature selection technique is applied to determine the best subset of features. Thirdly, an ensemble learning algorithm consisting of three individual learning algorithms namely Logistic Regression (LR), Support Vector Machine (SVM), and k-Nearest Neighbors (KNN) is used in the classification stage to predict conditions of PV systems based on a weighted voting approach. Moreover, we apply a genetic algorithm to optimize the weights assigned to the algorithms in order to detect electrical faults in PV systems with higher accuracy. The experimental results demonstrate the proposed model is very efficient and reliable to diagnose open-circuit (OC) and line-line (LL) faults in PV arrays under any challenging conditions with an accuracy of 99%. [ABSTRACT FROM AUTHOR]

• Development of a fault detection algorithm on aIRT images highlighting its main challenges, shortcomes, and workarounds; • Utilization of an orthomosaic reconstruction software package describing the challenges that the workflow impose for the automatic fault detection algorithms; • Evaluation of the impact of different flight configurations and dataset on the processing time and results; • Application of the proposed framework in real cases, evaluating the challenges imposed by real datasets. As large-scale Photovoltaic (PV) power plants are being expanded in installation number and capacity, aerial infrared thermography (aIRT) has proven to be effective in detecting at different phases of their development, construction and commissioning to operation and maintenance. However, evaluating the aerial imagery over hundreds of hectares fields of PV arrays is very time-consuming and subject to human error. This paper proposes a complete framework for automatically detecting faults in large-scale PV power plants and their physical location inside the plant site. To this end, a Mask-RCNN algorithm is developed and fine-tuned for instance segmentation using a dataset of 93 samples collected in an aIRT flight campaign in Brazil. The results are combined with orthomosaic techniques to create an orthomap of the PV system with the highlighted faults. The proposed method has been tested to automatically detect the faults in two power plants. Several tests were performed to improve the algorithm's accuracy, resulting in high-accuracy results for detecting and localizing hot spots in PV plants and disconnected substrings. The resulting maps could successfully show the location of these faults with high accuracy (10% of false positives). [ABSTRACT FROM AUTHOR]

Reliability assessment of Photovoltaic (PV) modules is very crucial to increase the service lifetime of PV systems. This study assesses the degradation rate and reliability of PV modules operated for twelve years under the tropical climatic condition in sub-Saharan. For this purpose, various characterization techniques, namely visual inspection, infrared (IR) thermography assessment, and current-voltage (I-V) characterization, have been employed to evaluate the performance of PV modules. Moreover, the functioning of bypass diodes has been tested under partial shading situations. The results demonstrate that after twelve years of outdoor operation, the short-circuit current (I sc) of modules have been degraded up to 16.4% with an average decrease of 11.7% compared to the nameplate values. The open-circuit voltages (V oc) were reduced from 11.4% to 17.1% with a mean of 14.8%. The decline in Fill Factor (FF) of the modules ranges from 11.3% to 24.2%, and the losses of power output were between 34.5% and 41.4%. Moreover, the visual and thermography assessment reveals that the PV modules are severely affected by various failures such as EVA browning, cell interconnects ribbons browning and the corrosion of solder bonds. The results show that averagely, the FF is the most significant factor influencing the loss in the power output of the modules. • Degradation of PV modules under hot, humid tropical climate for 12 years assessed. • PV module characteristics normalized at STC using derived translation equations. • Influence of module parameters on degradation evaluated using regression analysis. • The average degradation of nominal power is 3.19%/year. • The combination of Isc-Voc had the most impact on the variation of nominal power. [ABSTRACT FROM AUTHOR]

• A comprehensive review of LL fault diagnosis methods in PV systems. • An ensemble learning model based on probabilistic strategy is developed. • The main features of the fault are extracted from I-V curves using a Simulink based model of PV arrays. • A feature selection algorithm is used for each learning algorithm to reduce the dataset and increase the classification accuracy. • The proposed model is very reliable and able to detect and classify all the LL faults accurately. The fault diagnosis of photovoltaic (PV) arrays aims to increase the reliability and service life of PV systems. Line-Line (LL) faults may remain undetected under low mismatch level and high impedance due to low currents of faults, resulting in power losses and fire potential disaster. This paper proposes a novel and intelligent fault diagnosis method based on an ensemble learning model and Current-Voltage (I-V) characteristics to detect and classify LL faults at the DC side of PV systems. For this purpose, first, the key features are extracted via analyzing I-V characteristics under various LL fault events and normal operation. Second, a feature selection algorithm has been applied to select the best features for each learning algorithm in order to reduce the amount of data required for the learning process. Third, an ensemble learning model is developed that combines several learning algorithms based on the probabilistic strategy to achieve superior diagnostic performance. Here, we find an excellent agreement between simulation and experimental results that the proposed method can obtain higher accuracy in detecting and classifying the LL faults, even under low mismatch levels and high fault impedances. In addition, the comparison results demonstrate that the performance of the proposed method is better than individual machine learning algorithms, so that the proposed method precisely detects and classifies LL faults on PV systems under the different conditions with an average accuracy of 99% and 99.5%, respectively. [ABSTRACT FROM AUTHOR]

Line-line (LL) faults in photovoltaic (PV) arrays lead to a reduction in the output power and service life of PV systems, and if not detected and eliminated in time, may cause catastrophic fire hazards. The severity of LL faults could have an adverse effect on the performance of protection devices and depends on two main factors: 1) the number of modules involved in the fault, and 2) the resistivity of the fault path. In this study, first, LL faults are classified using a support vector machine (SVM) and then the main process of LL faults resistance (R f) estimation is carried out to determine the accurate severity of the fault. Conventionally, to measure the resistivity of the fault path, it is inevitable to employ multiple additional current and voltage sensors, which does not seem economically convenient and practical. Therefore, in this study, artificial intelligence (AI) has been utilized to estimate the fault resistance thus eliminating all the additional sensors. To this end, a regression-based dense neural network (R-DNN) is proposed and fine-tuned to predict the R f in a continuous interval of 0–24 Ω. Also, to provide the proposed model with a more organized thus simpler interpretation of the initial dataset, novel tangential features are constructed using five predefined points on the PV array current-voltage (I–V) characteristic curve. The experimental results show that the model is perfectly capable of diagnosing the LL faults and estimating the fault resistance values under various conditions. Finally, a comparison is carried out between the final R-DNN model and several regression-based machine learning algorithms which shows that the proposed R-DNN outperforms the other algorithms in terms of accuracy and execution time. [ABSTRACT FROM AUTHOR]

Early detection of faults in photovoltaic (PV) systems is crucial for improving their lifespan and reliability. Conventional protection devices may not be able to detect electrical faults under critical conditions. This usually occurs due to (i) the current-limiting nature and non-linear output characteristics of PV arrays, (ii) the changing environmental conditions (such as irradiance and temperature variations), and (iii) the influence of the maximum power point tracker (MPPT), particularly in the presence of critical fault impedance values or mismatch levels. Therefore, modern data-driven methods are required for fault detection and classification in PV arrays. However, careful investigation in previous studies reveals the existing gaps and limitations, such as poor accuracy, and incomprehensive models which have not considered various electrical faults, low mismatch levels, and critical fault impedance values. To this end, in this study, a comprehensive multilayer model is proposed which consists of six layers for diagnosis, classification and severity identification of electrical faults in PV arrays. Each layer adopts a weighted ensemble learning (WEL) algorithm consisting of three classifiers namely Support Vector Machine (SVM), Naïve Bayes (NB) and Logistic Regression (LR). Besides, a genetic algorithm (GA) is employed in each layer to find the optimal weights for each classifier (i.e., SVM, NB and LR). The initial dataset is acquired through an investigation into the PV array current–voltage (I-V) characteristic curve and extraction of several features under various faulty and normal conditions. To reduce the dataset dimensionality thus computationally simplifying the training process, the sequential floating forward selection (SFFS) algorithm is utilized in each layer as a powerful feature selection technique. The results show a highly accurate performance in fault diagnosis, classification, and severity assessment, with an average simulation and experimental accuracies of 98.9% and 98.37%, respectively. [ABSTRACT FROM AUTHOR]

Defective components within solar photovoltaic (PV) arrays overheat, resulting in particular temperature patterns under the long-wave thermal infrared (LWIR) spectrum. The detection and on-field localization of these patterns is of paramount aid to the operations and maintenance of PV installations. In this context, we develop a two-layer end-to-end inspection solution for the detection, quantification and on-field localization of overheated regions on PV arrays from LWIR UAV imagery. Layer 1 generates a georeferenced orthomosaic of the inspected site via a Structure from Motion-MultiView Stereo (SfM-MVS) photogrammetric acquisition/post-processing workflow. Layer 2 is a tile-based deep semantic segmentation stage that extracts and quantifies the affected regions from the generated orthomosaic. We collect aerial images from 103 PV sites, comprising approximately 342 000 modules. After a SfM-MVS workflow, we produce and annotate 7910 orthorectified unique affected tiles, posteriorly augmented to prepare the state-of-the-art dataset in terms of size and representativeness. Through a training/cross-validation and test process, we investigate the implementation of 9 models in the segmentation process: FCN, U-Net, FPN, DeepLab, LinkNet, DANet, CFNet, ACFNet and TransU-Net, each of which experimented with 2 backbones: ResNet50 and DenseNet121. The models feature efficient encoder-to-decoder feature map transfers, pyramidal feature recognition, spatial and channel attention, feature co-occurrence, class center as well as vision transformers. The best performance is achieved by FPN-DenseNet121, with a mean mIoU of 93.44% and an F1-score of 96.39% on our test set. The two-layer solution takes the best of the data-centric and model-centric paradigms, alongside addressing the limitations of conventional inspection procedures. It is put into a concrete application framework, where it provides a pixel-based and a tile-based quantification of the affected regions within a PV plant. The results are promising, and the selected model can be deployed efficiently for extensive aerial monitoring of large-scale PV plants. [ABSTRACT FROM AUTHOR]

The greenhouse industry is an energy-intensive sector with a heavy reliance on fossil fuels, contributing to substantial greenhouse gas (GHG) emissions. Addressing this issue, the employment of energy-saving strategies along with the replacement of conventional energy sources with renewable energies are among the most feasible solutions. Over the last few years, solar energy has demonstrated great potential for integration with agricultural greenhouses. The present study reviews the progress of solar greenhouses by investigating their integration with solar energy technologies including photovoltaic (PV), photovoltaic-thermal (PVT), and solar thermal collectors. From the literature, PV modules mounted on roofs or walls of greenhouses cause shading which can adversely affect the growing trend of cultivated crops inside. This issue can be addressed by using bifacial PV modules or employing sun trackers to create dynamic shades. PVT modules are more efficient in producing both heat and electricity, and less shading occurs when concentrating modules are employed. In terms of using solar thermal collectors, higher performance values have been reported for greenhouses installed in moderate climate conditions. Further, in this review, the employment of thermal energy storage (TES) units as crucial components for secure energy supply in solar greenhouses is studied. The usage of TES systems can increase the thermal performance of solar greenhouses by 29%. Additionally, the most common mathematical models utilized to describe the thermal behavior of solar greenhouses are presented and discussed. From the literature, machine learning algorithms have shown a better capability to describe the complex environment of greenhouses, but their main drawback is less reliability. Notwithstanding the progress which has been made, further improvements in technology and more reductions in costs are required to make the solar greenhouse technology a solution to achieve sustainable development. [ABSTRACT FROM AUTHOR]