Photovoltaic modules experience gradual degradation and sudden failures that reduce energy yield, reliability, and safety, motivating the use of imaging-based diagnostic techniques. Imaging approaches enable the detection of electrical and physical defects that are often invisible through conventional visual inspection. Studies have reported that microcracks, hotspots, potential-induced degradation, and light-induced degradation can noticeably affect PV module reliability. This review therefore examines both established field-deployed diagnostic techniques (technology readiness level > 5) and emerging approaches that are under development (technology readiness level <5). Most existing studies focus on the development or evaluation of individual imaging techniques, while some review articles discuss multiple methods without providing detailed comparisons between conventional and emerging diagnostic approaches across different fault types. As a result, a systematic comparison of the diagnostic capabilities of the available imaging modalities remains limited. To address this gap, this review presents a structured study and fault-centric benchmarking of various imaging-based PV inspection techniques, emphasizing fault visibility and diagnostic relevance across imaging modalities rather than relying solely on reported accuracy metrics. Additionally, a hybrid scope–mapping systematic review methodology is applied, in which peer-reviewed studies are screened, classified, and synthesized based on fault type and technological maturity. Based on results reported in the literature, machine learning-assisted infrared thermography has achieved detection accuracies of 94–98%, while deep learning-based electroluminescence methods have reported accuracies of up to 97.8%. Ultraviolet fluorescence techniques have demonstrated crack detection rates exceeding 91% and inspection throughput up to 10–15 times higher than near-infrared inspection under comparable operating conditions. These performance values originate from different studies, datasets, and experimental conditions and are therefore intended to illustrate representative capabilities rather than enable direct comparison between techniques. Emerging approaches such as daylight luminescence and magnetic-field-based diagnostics are also gaining attention, although their broader use remains limited by operational complexity and signal-to-noise challenges.
This paper reports a study on the structural, magnetic and magnetotransport properties of the mixed-valence perovskite manganite La0.7Ca0.18Ba0.12Mn0.95Sn0.05O3 (LCBMSO), synthesized by a solid-state reaction method. The results of X-ray powder diffraction analysis confirm that the sample possesses a single orthorhombic phase with space group Pnma. The R,ietveld refinement results reveals that LCBMSO contains distorted MnO6 octahedron. Micrographs obtained by scanning electron microscopy showed that the sample grains have a polygonal shape and are in the micrometer size range. Fourier transform infrared spectroscopy analysis confirms the presence of Mn-O-Mn and Mn-O stretching vibration. The magnetization-temperature curve displays a paramagnetic-ferromagnetic transition at TC = 145 K. A slight bifurcation between the zero-field curve and the cooling-field curve was noticed, which is attributed to spin-glass behavior. Based on the hysteresis cycle, a soft ferromagnetic behavior was observed in our sample at temperatures of 1.8 and 100 K. The electrical resistivity vs temperature curve shows a metal-insulator transition at TMI = 154 K. The magnetoresistance ratio reached 30% at an applied magnetic field of 1 T, making the LCBMSO material an attractive candidate for use as a magnetoresistive sensor in various industrial applications. The temperature coefficient of resistivity reached 3.35%, making this material suitable for use in infrared and bolometric detectors. The relation rho = rho 0-rho 0.5T0.5 + rho 2T2 + rho 5T5 was employed to fit the low-temperature resistivity data below TMI, whereas the variable range hopping and small polaron hopping models were used to fit the data in the insulating region above TMI.
This paper presents Refactal, an online serious game developed to introduce software engineering students to design patterns, code smells, and SOLID principles. Structured as an escape room, the game features progressive levels and integrates refactoring-based problem solving. The game's design and a pilot exploratory evaluation are reported. Over three iterative testing phases, involving 33, 16, and 10 computer science students, data were collected through user experience questionnaires, in-game performance metrics and, during the final phase, pre/post-tests assessing conceptual learning. Results indicate that students perceived Refactal positively, particularly regarding satisfaction, scenario relevance, and overall experience. Preliminary evidence of learning gains, especially among novices, suggests that Refactal may support the acquisition of advanced OOP principles. While the small sample size limits generalization, these findings highlight Refactal's potential and motivate future controlled studies with larger and more diverse cohorts.
Enhancing the cooling of electronic components enables the development and downsizing of electronic devices. This study investigates the impact of inclined sidewalls on cooling performance in 3D turbulent mixed convection in a ventilated parallelepiped cavity with three heat sources on the bottom wall. The enclosure's sidewalls contain slots that allow airflow in and out to dissipate the generated heat. The finite volume method is used to solve the equations governing the configuration considered. Comparisons with available experimental data and current simulation results show good agreement. The flow and thermal fields are analyzed using particle trajectories, iso-surfaces, and the average Nusselt number. Various airflow slot positions are tested. Moreover, the vertical walls are inclined at different angles, positive (alpha = 5 degrees, 10 degrees, 15 degrees, and 20 degrees) and negative (alpha = -5 degrees, 10 degrees, 15 degrees, and -20 degrees), to determine the most appropriate position for cooling the heat sources. It is found that the critical position of airflow slots in convective cooling of heat sources corresponds to Case 4 (LD-RD). Furthermore, inclined vertical walls at positive angles have a greater effect on cooling heat sources than those at negative angles. The results provide new insights into the cooling process of electronic components.
Aim of the study The objective of this study is to evaluate the impact of three kernel functions-Pearson VII, radial basis function (RBF), and polynomial-on the predictive performance of Support Vector Regression (SVR) and Gaussian Process Regression (GPR) models. Materials and methods Three machine learning models-Random Forest (RF), Support Vector Regression (SVR), and Gaussian Process Regression (GPR)-were applied to estimate monthly evaporation at Boukourdane Dam, Algeria. The dataset included 240 observations over 20 years, with the following inputs: max./min. air temperature, relative humidity, wind speed, and water temperature; the output being: evaporation. Results and conclusions Model performance was evaluated via Correlation Coefficient (CC), Root Mean Square Error (RMSE), and Mean Absolute Error (MAE). RF outperformed GPR and SVR across kernels, achieving MAE = 1.01 mm, RMSE = 1.29 mm, and CC = 0.81 in testing. Moreover, the Pearson VII kernel delivered the highest accuracy within both the GP and SVM frameworks. Sensitivity analysis highlighted relative humidity as the most influential factor in evaporation forecasting.
