Multimodal Foundation Model (MFM), like ChatGPT and Gemini, have emerged as powerful tools for their exceptional natural language processing capabilities and their emerging potential in image analysis. This paper investigates the application of MFMs for photovoltaic (PV) fault detection through image analysis, focusing on ChatGPT 4.0 and Gemini 1.5 Pro. Three types of PV images and the corresponding common PV faults are detected: bird droppings using visible images, cell cracks via electroluminescence (EL) images, and hotspots using infrared (IR) images. Among the two models, Gemini 1.5 Pro demonstrated superior performance, achieving near-perfect results with an average F1 score of 0.97, consistently outperforming ChatGPT 4.0 in accuracy and reliability. Unlike traditional machine learning (ML) models, MFMs can operate in a zero shot manner that does not require additional training by the user, and the input images are not limited by size, angle, scope, or PV technology. The strong adaptability and user-friendliness make MFM a promising tool for analyzing PV images and advancing health monitoring for PV modules.

Solar modules in utility-scale systems are expected to maintain decades of lifetime to rival conventional energy sources. However, cyclic thermomechanical loading often degrades their long-term performance, highlighting the importance of effective design to mitigate thermal expansion mismatches between module materials. Given the complex composition of solar modules, isolating the impact of individual components on overall durability remains a challenging task. In this work, we analyze a comprehensive data set that comprises bill-of-materials (BOM) and thermal cycling power loss from 251 distinct module designs to identify the predominant design factors and their impacts on the thermomechanical durability of modules. The methodology of our analysis combines machine learning modeling (random forest) and Shapley additive explanation (SHAP) to correlate design factors with power loss and interpret the model's decision-making. The interpretation reveals that silicon type (monocrystalline or polycrystalline), encapsulant thickness, busbar numbers, and wafer thickness predominantly influence the degradation. With lower power loss of around 0.6% on average in the SHAP analysis, monocrystalline cells present better durability than polycrystalline cells. This finding is further substantiated by statistical testing on our raw data set. The SHAP analysis also demonstrates that while thicker encapsulants lead to reduced power loss, further increasing their thickness over around 0.6 to 0.7 mm does not yield additional benefits, particularly for the front side one. In addition, other important BOM features such as the number of busbars are analyzed. This study provides a blueprint for utilizing explainable machine learning techniques in a complex material system and can potentially guide future research on optimizing the design of solar modules.

To realize ultrafast supercapacitors, a series of ultrahigh-rate (>= 1000 mV s(-1)) anodes that break a well-known "energy vs. power dilemma" in aqueous electrolytes have been successfully developed over the past five years. However, their matching cathodes are still limited by slow ion transport dynamics and oxidation. Here, we report a series of hydrated films that comprise small sheets of reduced graphene oxide (SSs-rGO, <100 nm in average lateral size) and feature short interlayered pathways (similar to 100 nm) for rapid ion transport and high oxidation resistance. As the first ultrahigh-rate cathode with areal capacitance (C-a) satisfying an industrial requirement (>0.6 F cm(-2)), the SSs-rGO electrode (5.0 mg cm(-2)) delivers a C-a of 0.61 F cm(-2) and a gravimetric capacitance of 123 F g(-1) at an ultrahigh-rate of 3000 mV s(-1). Combining with a Ti3C2Tx anode, the cathode enables a 1.8 V ultrafast aqueous supercapacitor that delivers energy densities of 0.14 and 0.09 mWh cm(-2) for discharges in 1.79 and 0.97 s (similar to 2000 and 3700 C rate), respectively. These values double (at 2000 C) and almost ten-fold (at 3700 C) those of the ever-reported supercapacitors operating at the corresponding rates. The present strategy paves a road to ultrafast (>1000 C) and high-energy-density supercapacitors, by which energy charge/discharge can finish within 3.6 s.

Energy densities of widely used carbon-based supercapacitors can be enhanced by expanding carbon's electrochemical stability potential windows (ESPWs) or by introducing pseudocapacitance to the energy storage/release processes. Nevertheless, maintaining materials' rate performance and cycling stability after these treatments remains challenging. Here, we apply a cyclic voltammetry treatment with various maximum oxidation potentials (MOPs) to introduce oxygenated functionalities onto the electrochemically active surfaces of carbon materials (e. g., activated carbon, graphene, and carbon nanotubes). This treatment expands the carbons' ESPWs by improving oxidation resistance and introduces highly stable proton-involved pseudocapacitance on carbons. By applying moderate MOPs of +0.6 V and +0.7 V (vs. Hg/Hg2SO4) in 3 M H2SO4, the ESPWs of the activated carbon electrodes (5.0 mg cm- 2 ) are extended from 1.0 V to 1.2 V and 1.3 V, respectively. Meanwhile, their capacitances increase from 271.5 F g- 1 to 369.2 F g- 1 and 395.0 F g- 1 at a low rate of 10 mV s-1 (from 304.7 F g- 1 to 421.9 F g- 1 and 472.4 F g- 1 at 1 A g- 1 ), and from 186.8 F g- 1 to 243.7 F g- 1 and 217.8 F g- 1 at an ultrahigh rate of 1000 mV s-1, respectively. Thus, the energy densities of corresponding symmetric devices increase by 76.0 % and 101.3 % at 1 A g-1, and by 110.7 % and 27.2 % at 100 A g-1, respectively. This work provides a simple and facile way to significantly optimize the active materials for energy storage, which is expected to be extended to other electrode systems for supercapacitors in the near future.

Photovoltaic (PV) systems have become a cornerstone of renewable energy strategies, particularly due to the significant reduction in solar power costs over the past decade. However, the long-term reliability of PV installations presents a persistent challenge, requiring the development of advanced monitoring and predictive maintenance strategies. A wide range of data types is used to evaluate the health of PV systems, including environmental conditions, electrical performance, and inspection imagery. These data enable methodologies such as machine learning (ML) models for lifetime prediction and computer vision techniques for defect detection. However, the acquisition of high-quality and comprehensive data is difficult, particularly in terms of long-term consistency and data variety. Publicly available data sets serve as valuable resources for addressing these challenges, but they often suffer from fragmentation and are difficult to access. This paper presents a comprehensive review of existing open-source data sets related to PV degradation, analyzing their features, functionalities, and potential applications. We categorize these data sets based on the specific aspects of PV system information they cover, such as environmental conditions, operational monitoring, image inspection and module materials, and propose relevant tools and ML models for processing them. In addition, we propose practices for future data collection and usage, while also discussing potential directions in data-driven research. Our aim is to enhance data utilization and publication among researchers and industry professionals, promoting a deeper understanding of the role of data in enhancing the performance and durability of PV systems.

To enable health monitoring and fault diagnosis of PV modules using current-voltage characteristics (I-V curves), it is generally necessary to correct the I-V curves measured under different environmental conditions to the standard condition. The most common correction methods are those from IEC 60891: 2021 standard. However, these methods can introduce significant errors when dealing with degraded PV modules due to the inability to account for changes in resistance. To address this, we propose an improved I-V curve procedure, denoted Pdynamic, which considers different types of degradation by dynamically deriving the correction coefficients from the measured I-V curves. To evaluate the performance, we simulate I-V curves across a wide range of irradiance and temperature for the healthy and degraded module, where the degradation involves increased series resistance, decreased shunt resistance, or both. The results reveal that Pdynamic can produce corrected I-V curves closer to the reference ones than Procedures 1, 2, and 4 of the IEC 60891:2021 standard. Moreover, Pdynamic exhibits resilience to both seasonal fluctuations and varying levels of degradation. These results highlight Pdynamic as a promising and robust I-V curve correction method, particularly for degraded PV modules. A Python-based open-source tool for this procedure is also available at https://github.com/DuraMAT/IVcorrection.

Constructing a LiF-rich solid electrolyte interphase (SEI) is a feasible strategy for inhibiting lithium (Li) dendrites of Li metal anodes (LMAs). However, selecting appropriate F-containing additives with efficient LiF contribution is still under active research. Herein, a series of fluorinated additives with diverse F/C molar ratios are investigated, and we demonstrate that the hexafluoroglutaric anhydride (F6-0) holds the best capability to derive the LiF-rich SEI in regular carbonate electrolytes (RCEs). To ameliorate the decomposition kinetics of the F6-0, LiNO3 (LNO) as an adjuvant is further introduced in the system. As a result, the reduction efficiency of F6-0 is increased to 91% under the F6-0/LNO synergistic effect, enabling the LMA with a uniform LiF-rich SEI in the RCE with merely 4 vol. % F6-0/LNO (F6L) addition. The LiNi0.8Co0.1Mn0.1O2||Li-20μm full-cell with the F6L also showcases better cycling and rate performances than the cases with other F-containing additives.

A biotopologically structured configuration constructed by an energy-saving electrodeposition method for low-temperature supercapacitors.

Power prediction is crucial to the efficiency and reliability of Photovoltaic (PV) systems. For the model-chain-based (also named indirect or physical) power prediction, the conversion of ground environmental data (plane-of-array irradiance and module temperature) to the output power is a fundamental step, commonly accomplished through physical modeling. The core of the physical model lies in the parameters. However, traditional parameter estimation either relies on datasheet information that cannot reflect the system's current health status or necessitates additional I-V characterization of the entire array, which is not commonly available. To address this, our paper introduces PVPro, a dynamic physical modeling method for irradiance-to-power conversion. It extracts model parameters from the recent production data without requiring I-V curve measurements. This dynamic model, periodically-updated (as short as daily), can closely capture the actual health status, enabling precise power estimation. To evaluate the performance, PVPro is compared with the smart persistence, nominal physical, and various machine learning models for day-ahead power prediction. The results indicate that PVPro achieves an outstanding power estimation performance with the average nMAE =1.4 the error by 17.6 demonstrates robustness across different seasons and weather conditions. More importantly, PVPro can also perform well with a limited amount of historical production data (3 days), rendering it applicable for new PV systems. The tool is available as a Python package at: https://github.com/DuraMAT/pvpro.

Lithium-ion batteries (LIBs) are the cornerstone of the transition to renewable energy and can power a wide range of devices such as smartphones as well as electric vehicles, although they face some major challenges such as resource scarcity, safety risks and thermal runaway. This paper provides an insightful discussion on the mechanism of operation of LIBs, their applications, and its limitations, emphasizing that LIBs, while widely used in electric vehicles and portable electronics, have certain limitations, like thermal instability. The paper also explores new alternatives, including solid-state and lithium-metal batteries, which have higher energy densities but also suffer from problems such as dendrite growth, for which it cites improvement strategies, such as innovative cathode and negative electrode materials and electrolyte additives, with which to increase the lifetime, safety and performance of the batteries. At the same time, the paper also considers the issue of resource constraints and proposes a scalable and sustainable approach to energy storage needs with a multi-pronged strategy necessary to overcome existing barriers.

Ionic conductors have great potential for interesting tunable physical properties via ionic liquid gating and novel energy storage applications such as all-solid-state lithium batteries. In particular, low migration barriers and high hopping attempt frequency are the keys to achieve fast ion diffusion in solids. Taking advantage of the oxygen-vacancy channel in Li_xSr_2Co_2O_5, we show that migration barriers of lithium ion are as small as 0.28 0.17eV depending on the lithium concentration rates. Our first-principles calculation also investigated hopping attempt frequency and concluded the room temperature ionic diffusivity and ion conductivity is high as 10^-7∼10^-6 cm^2 s^-1 and 10^-3∼10^-2 S·cm^-1 respectively, which outperform most of perovskite-type, garnet-type and sulfide Li-ion solid-state electrolytes. This work proves Li_xSr_2Co_2O_5 as a promising super-ionic conductor.

Physics-based circuit parameters like series and shunt resistance are essential to provide insights into the degradation status of photovoltaic (PV) arrays. However, calculating these parameters typically requires a full current-voltage characteristic (I -V curve), the acquisition of which involves specific measurement devices and costly methods. Thus, I -V curves of the PV system level are often not available. This paper proposes a meth-odology (PVPRO) to estimate these I -V curve parameters using only operation (string-level DC voltage and current) and weather data (irradiance and temperature). PVPRO first performs multi-stage data pre-processing to remove noisy data. Next, the time-series DC data are used to fit an equivalent circuit single-diode model (SDM) to estimate the circuit parameters by minimizing the differences between the measured and estimated values. In this way, the time evolutions of the SDM parameters are obtained. We evaluate PVPRO on synthetic datasets and find an excellent estimation of both SDM and the key I-V parameters (e.g., open-circuit voltage, short-circuit current, maximum power, etc.) with an average relative error of 0.55%. The performance, especially the extracted degradation rate of parameters, is robust to various measurement noises and the presence of faults. In addition, PVPRO is applied to a 271 kW PV field system. The relative error between the real and estimated operation voltage and current is less than 1%, suggesting that degradation trends are well captured. PVPRO represents a promising open-source tool to extract the time-series degradation trends of key PV parameters from routine operation data.

The effect of cracks in solar cells on the long-term degradation of photovoltaic (PV) modules remains to be determined. To investigate this effect in future studies, it is necessary to quantitatively describe the crack features (e.g., length) and correlate them with module power loss. Electroluminescence (EL) imaging is a common technique for identifying cracks. However, it is currently challenging and time-consuming to identify cracks in a large number of EL images and quantify complex crack features by human inspection. This article introduces a fast semantic segmentation method ($\sim$0.18 s/cell) to automatically segment cracks from EL images and algorithms to extract crack features. We fine-tuned a UNet neural network model using pretrained VGG16 as the encoder and obtained an average F1 score of 0.875 and an intersection over union score of 0.782 on the testing set. With cracks and busbars segmented, we developed algorithms for extracting crack features, including the crack-isolated area, the brightness inside the isolated area, and the crack length. We also developed an automatic preprocessing tool for cropping individual cell images from EL images of PV modules ($\sim$0.72 s/module). Our codes are published as open-source an software, and our annotated dataset composed of various types of cells is published as a benchmark for crack segmentation in EL images.

Photovoltaic (PV) systems can operate off the maximum power point (MPP) for various reasons. Understanding when off- MPP behavior occurs is essential to the maintenance and operation (O&M) of PV systems. To detect off-MPP data, a reference power is usually needed, which can be obtained by system modeling that generally relies on physical model parameters. Traditional methods commonly obtain these parameters based on the initial condition of the PV system such as from the module datasheet. However, these parameters often do not reflect the current condition of the on-site PV system, which is likely to suffer from degradation and faults after years of operation with degraded parameters. Thus, we propose an off-MPP analysis algorithm based on the PV-Pro method, which can extract the model parameters (like series and shunt resistance) at the current operating condition only using the routine production data. In this way, the system power, current, and voltage can be accurately modeled. The off-MPP points are detected by comparing the measured power with the one modeled by PV-Pro. Points with large disagreement in power are further analyzed by deconvolving it into the error of the current and voltage at MPP, which allows tracing the error source of the off- MPP and provides valuable information for the O&M of PV systems. This off-MPP analysis is demonstrated on a 271kW PV field system, where it is shown that most of the off- MPP points are caused by the reduced DC current.

This study compared module power loss for 36 modules that endured various accelerated aging test sequences before installation outdoors on a 10-kWp array in Birmingham, AL, USA for 1.72 to 2.72 years. Twelve modules endured standard IEC 61215 aging tests and 24 endured Qualification Plus (Qual Plus). Modules in each group were further split into two test sequences with different exposures. Electrical parameter variations were analyzed as a function of aging test and field exposure history. Fill factor loss was determined to be the cause of observed decreases in power output during accelerated aging tests, while decreases in both open circuit voltage and fill factor dominated the power loss during subsequent on-sun testing. Quantified cell crack features were extracted via computer vision tools from electroluminescence images and correlated with power loss. Results illustrate that standard aging tests led to negligible cracks, while Qual Plus test sequences yielded more severe cracks. While correlating results from qualification tests with in-field performance degradation parameters remains a challenge, this study provides new insights on specific environmental stressors and crack features that may play a role in power loss. Insights on accelerated aging protocols are discussed.

Cracks on solar cells can cause degradation of photovoltaic (PV) modules, and electroluminescence (EL) images are a common technique for identifying cracks. However, to process a large number of such images it is necessary to develop automated routines for analysis. This article introduces a fast semantic segmentation method (~0.18s/cell) to segment crack lines, cross cracks, busbars, and dark areas on EL images of PV modules. We trained a UN et neural network model on a training set of 1,272 images, and we evaluated its performance on a validation set of 206 images and a testing set of 359 images. We report the performance on the testing set with an average F1 score of 0.875 and an IoU score of 0.782. We introduce our algorithm of predicting the worst-case degradation area with cracks detected. We also demonstrate our automatic preprocessing tool of cropping individual cell images from EL images of PV modules (~0. 72s/module). Our methods are published as open-source software and might be used to segment other kinds of defects or similar types of images by transfer learning in the future.

Physics-based circuit parameters like series and shunt resistance are also essential to provide insights into the degradation modes of PV arrays. However, the calculation of these parameters typically relies on I-V curves, which require specific measurement devices and may impede the regular operation of PV systems. Although a reference module equipped with I-V tracers may be installed next to the array, the recorded I-V curves cannot fully represent the array condition. Therefore, I-V curves of the entire array are not always available to obtain the circuit parameters, especially for large-scale PV power plants. Therefore, this paper proposes a methodology (PVPRO) to estimate these parameters for degradation analysis without the need for I-V tracers. Rather, it estimates these parameters using only operation (DC voltage and current) and weather data (irradiance and module temperature). PVPRO first performs a multi-stage data cleaning and filtering. Next, the time-series DC data are clipped into windows to fit the equivalent single-diode model to estimate the circuit parameters on minimizing the loss between real and estimated values. Finally, the estimated parameters are plotted as function of time to quantitatively analyze the degradation trends. PVPRO is evaluated on synthetic datasets in the presence of noise and is applied to a well-qualified field system (2015-2019). The degradation trend by PVPRO is in reasonable agreement with the ground truth, especially for the seasonal variation (correlation coefficient of 0.71). Future work will apply PVPRO to more large-scale PV systems and deconvolve the degradation pathways. The current version of PVPRO is available on Github: https://github.com/DuraMAT/pvpro

Solar photovoltaic (PV) modules are susceptible to manufacturing defects, mishandling problems or extreme weather events that can limit energy production or cause early device failure. Trained professionals use electroluminescence (EL) images to identify defects in modules, however, field surveys or inline image acquisition can generate millions of EL images, which are infeasible to analyze by rote inspection. We develop a rapid automatic computer vision pipeline (∼0.5 seconds/module) to analyze EL images and identify defects including cracks, intra-cell defects, oxygen-induced defects, and solder disconnections. Defect identification is achieved with a machine learning model (Random Forest, ResNet models and YOLO) trained on 762 manually-labeled EL images of PV modules. We compare model performance on an imbalanced real-world validation set containing 134 EL images and determine that ResNet18 and YOLO are the optimal models; we next evaluated these models on a dedicated testing set (129 module images) with resulting macro F1 scores of 0.83 (ResNet18) and 0.78 (YOLO). Using a field EL survey of a PV power plant damaged in a vegetation fire, we analyze 18,954 EL images (2.4 million cells) and inspect the spatial distribution of defects on the solar modules. The results find increased frequency of ‘crack’, ‘solder’ and ‘intra-cell’ defects on the edges of the solar module closest to the ground after fire. We also find an abnormal increase of striation rings on cells which were assumed to be caused mainly in fabrication process. Our methods are published as open-source software. It can also be used to identify other kinds of defects or process different types of solar cells with minor modification on models by transfer learning.

Compared to traditional optical microscope, lensless inline holographic microscope (LIHM) is more compact and low-cost. However, its resolution and imaging contrast are generally inferior mainly because of the twin-image background. In this paper we propose a deep learning-based approach to reduce the noise and enhance the imaging quality in LIHM by inter-modality learning from the traditional microscope images. By exploiting the denoising model in the learning processing, our network can be trained with a dataset synthesized from the direct-reconstructed images of LIHM and the high-resolution ground truth images obtained with a microscope. In the imaging process, other direct-reconstructed images of LIHM can then be enhanced by the trained denoising network. The image enhancement capability of our method was demonstrated by experiments with a U.S. Air force (USAF) target and a pumpkin stem sample. The results show that both the resolution and imaging contrast were significantly improved compared with traditional reconstruction methods in LIHM.