Classification machine learning models require high-quality labeled datasets for training. Among the most useful datasets for photovoltaic array fault detection and diagnosis are module or string current-voltage (IV) curves. Unfortunately, such datasets are rarely collected due to the cost of high fidelity monitoring, and the data that is available is generally not ideal, often consisting of unbalanced classes, noisy data due to environmental conditions, and few samples. In this paper, we propose an alternate approach that utilizes physics-based simulations of string-level IV curves as a fully synthetic training corpus that is independent of the test dataset. In our example, the training corpus consists of baseline (no fault), partial soiling, and cell crack system modes. The training corpus is used to train a 1D convolutional neural network (CNN) for failure classification. The approach is validated by comparing the model’s ability to classify failures detected on a real, measured IV curve testing corpus obtained from laboratory and field experiments. Results obtained using a fully synthetic training dataset achieve identical accuracy to those obtained with use of a measured training dataset. When evaluating the measured data’s test split, a 100% accuracy was found both when using simulations or measured data as the training corpus. When evaluating all of the measured data, a 96% accuracy was found when using a fully synthetic training dataset. The use of physics-based modeling results as a training corpus for failure detection and classification has many advantages for implementation as each PV system is configured differently, and it would be nearly impossible to train using labeled measured data.

Accurate diagnosis of failures is critical for meeting photovoltaic (PV) performance objectives and avoiding safety concerns. This analysis focuses on the classification of field-collected string-level current-voltage (IV) curves representing baseline, partial soiling, and cracked failure modes. Specifically, multiple neural network-based architectures (including convolutional and long short-term memory) are evaluated using domain-informed parameters across different portions of the IV curve and a range of irradiance thresholds. The analysis identified two models that were able to accurately classify the relatively small dataset (~400 samples) at a high accuracy (99%+). Findings also indicate optimal irradiance thresholds and opportunities for improvements in classification activities by focusing on portions of the IV curve. Such advancements are critical for expanding accurate classification of PV faults, especially for those with low power loss (e.g., cracked cells) or visibly similar IV curve profiles.

Failure detection methods are of significant interest for photovoltaic (PV) site operators to help reduce gaps between expected and observed energy generation. Current approaches for field-based fault detection, however, rely on multiple data inputs and can suffer from interpretability issues. In contrast, this work offers an unsupervised statistical approach that leverages hidden Markov models (HMM) to identify failures occurring at PV sites. Using performance index data from 104 sites across the United States, individual PV-HMM models are trained and evaluated for failure detection and transition probabilities. This analysis indicates that the trained PV-HMM models have the highest probability of remaining in their current state (87.1% to 93.5%), whereas the transition probability from normal to failure (6.5%) is lower than the transition from failure to normal (12.9%) states. A comparison of these patterns using both threshold levels and operations and maintenance (O&M) tickets indicate high precision rates of PV-HMMs (median = 82.4%) across all of the sites. Although additional work is needed to assess sensitivities, the PV-HMM methodology demonstrates significant potential for real-time failure detection as well as extensions into predictive maintenance capabilities for PV.

Although unique expected energy models can be generated for a given photovoltaic (PV) site, a standardized model is also needed to facilitate performance comparisons across fleets. Current standardized expected energy models for PV work well with sparse data, but they have demonstrated significant over-estimations, which impacts accurate diagnoses of field operations and maintenance issues. This research addresses this issue by using machine learning to develop a data-driven expected energy model that can more accurately generate inferences for energy production of PV systems. Irradiance and system capacity information was used from 172 sites across the United States to train a series of models using Lasso linear regression. The trained models generally perform better than the commonly used expected energy model from international standard (IEC 61724-1), with the two highest performing models ranging in model complexity from a third-order polynomial with 10 parameters ($R_{adj}^2$ = 0.994) to a simpler, second-order polynomial with 4 parameters ($R_{adj}^2=0.993$), the latter of which is subject to further evaluation. Subsequently, the trained models provide a more robust basis for identifying potential energy anomalies for operations and maintenance activities as well as informing planning-related financial assessments. We conclude with directions for future research, such as using splines to improve model continuity and better capture systems with low (≤1000 kW DC) capacity.

Sampling is an important step in the machine learning process because it prioritizes samples that help the model best summarize the important concepts required for the task at hand. The process of determining the best sampling method has been rarely studied in the context of graph neural networks. In this paper, we evaluate multiple sampling methods (i.e., ascending and descending) that sample based off different definitions of centrality (i.e., Voterank, Pagerank, degree) to observe its relation with network topology. We find that no sampling method is superior across all network topologies. Additionally, we find situations where ascending sampling provides better classification scores, showing the strength of weak ties. Two strategies are then created to predict the best sampling method, one that observes the homogeneous connectivity of the nodes, and one that observes the network topology. In both methods, we are able to evaluate the best sampling direction consistently.

The purpose of pvOps is to support empirical evaluations of data collected in the field related to the operations and maintenance (O&M) of photovoltaic (PV) power plants. pvOps presently contains modules that address the diversity of field data, including text-based maintenance logs, current-voltage (IV) curves, and timeseries of production information. The package functions leverage machine learning, visualization, and other techniques to enable cleaning, processing, and fusion of these datasets. These capabilities are intended to facilitate easier evaluation of field patterns and extraction of relevant insights to support reliability-related decision-making for PV sites. The open-source code, examples, and instructions for installing the package through PyPI can be accessed through the GitHub repository.