The meticulous monitoring and diagnosis of faults in photovoltaic (PV) systems enhances their reliability and facilitates a smooth transition to sustainable energy. This paper introduces a novel application of deep learning for fault detection and diagnosis in PV systems, employing a three-step approach. Firstly, a robust PV model is developed and fine-tuned using a heuristic optimization approach. Secondly, a comprehensive database is constructed, incorporating PV model data alongside monitored module temperature and solar irradiance for both healthy and faulty operation conditions. Lastly, fault classification utilizes features extracted from a combination consisting of a Convolutional Neural Network (CNN) and Bidirectional Gated Recurrent Unit (Bi-GRU). The amalgamation of parallel and sequential processing enables the neural network to leverage the strengths of both convolutional and recurrent layers concurrently, facilitating effective fault detection and diagnosis. The results affirm the proposed technique’s efficacy in detecting and classifying various PV fault types, such as open circuits, short circuits, and partial shading. Furthermore, this work underscores the significance of dividing fault detection and diagnosis into two distinct steps rather than employing deep learning neural networks to determine fault types directly.
To ensure the optimization of the energy generated by grid connected PV (photovoltaic) systems is necessary to plan a strategy of automatic fault detection. The analysis of current and voltage indicators have demonstrated effectiveness in the detection of permanent faults in the PV array in real time as short-circuits or open circuits present in the system. In this paper, the analysis of the evolution of these indicators is focused on the detection of temporary faults due to partial shade on the PV array or disconnection of the inverter in case of grid fluctuations of voltage or frequency to prevent islanding. These situations can be identified by observation of the evolution of both indicators and power losses due to these effects can be evaluated from them. The analysis and experimental validation were carried out in two grid connected PV systems in Spain and Algeria.Peer ReviewedPostprint (author's final draft
Photovoltaic (PV) systems based on multistring configuration are the best effective solution, given its advantages in terms of system availability, reliability, and energy efficiency. In this particular configuration each substring has its own dc–dc converter and a dedicated maximum power search algorithm which increase the cost and complexity. In this article, an efficient centralized global maximum power tracking (GMPPT) algorithm for multistring PV array subject to partial shading conditions is proposed. The algorithm is based on artificial bee colony (ABC) as an optimization approach to provide the optimal duty cycles allowing the extraction of the optimal global maximum power from each substring. In particular, the proposed approach allows significant reduction of the required sensors to only one pair of current and voltage sensors, at the common point of connection of the overall PV strings. The simulation study has been carried out under Cadence/Pspice and MATLAB/Simulink platforms on the I–V curves to confirm the effectiveness of the proposed algorithm when several shading patterns occur. In addition, complex shading pattern of a daily profile has been also carried out to demonstrate the GMPPT finding in dynamically variable conditions. Performance comparison against particle swarm optimization based maximum power point tracking algorithm and the traditional perturb and observe method has also been carried out. The obtained simulation and experimental results have shown the effectiveness and a good tracking capability of the proposed ABC algorithm in a multistring PV array configuration under uniform and nonuniform irradiance.
Floating photovoltaics (FPVs) provide various benefits especially where land is scarce (e.g., reducing land occupancy, water evaporation and environment control…), or when they are combined with hydropower plants (enhanced capacity factor and green energy generation). Software such as PV∗SOL, SAM and PVSyst® are commonly used for the design and simulation of land-based photovoltaic (PV) systems. However, when it comes to the simulation of photovoltaics installed on water surface, such software does not offer the option to directly simulate FPV systems.