Neural Network-Based Classification of String-Level IV Curves From Physically-Induced Failures of Photovoltaic Modules

Abstract

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.

A comprehensive review of unmanned aerial vehicle-based approaches to support photovoltaic plant diagnosis

“…Electrical data characterisation methods include I–V curve analysis [ 93 , 94 ], signal analysis, circuit and simulation models [ 6 ]. Current trends and innovations in today's O&M market include the application of ML techniques [ 38 , 43 , 63 , 85 ] and data-driven decision making processes [ [95] , [96] , [97] , [98] ].…”

Application of Artificial Neural Networks to photovoltaic fault detection and diagnosis: A review

“…Chen et al [97] combine I-V curve data with G and TM to form up a 404 feature matrix. Similar I-V curve-based approaches are also applied in [98,99]. Besides, Aziz et al adopted Continuous Wavelet Transform (CWT) [100] to generate scalograms (2-D graphs) from environmental and array electrical parameters.…”

Classification of Photovoltaic Failures with Hidden Markov Modeling, an Unsupervised Statistical Approach

“…Given the notable impact that reduced production can have on overall PV site revenue [2], significant attention has been given to failure detection and diagnosis methods to improve the reliability of these systems and reduce system downtime. Diverse types of data have been used to characterize failures, including current-voltage traces [7], electroluminescence/infrared imagery data [8], and maximum power points [9]. Researchers have also evaluated operations and maintenance (O&M) tickets to understand seasonal patterns in failure frequencies to inform operators of common failure modes and support planning efforts [5].…”