Document Type : Research Article
Authors
1
Department of Electrical Engineering, Laboratory of Electrical System, University of Sciences and Technology Houari Boumediene, USTHB, B. P. 32 El Alia 16111, Bab Ezzouar, Algiers, Algeria.
2
aDepartment of Electrical Engineering, Laboratory of Electrical System, University of Sciences and Technology Houari Boumediene, USTHB, B. P. 32 El Alia 16111, Bab Ezzouar, Algiers, Algeria
3
Centre de Développement des Energies Renouvelables, CDER, B. P. 62 Route de l’Observatoire Bouzareah 16340, Algiers, Algeria
4
cDepartment of Engineering, Faculty of Environment, Science and Economy, University of Exeter, Penryn Campus, Cornwall TR10 9FE, UK.
10.22059/jser.2026.422204.1782
Abstract
Accurate parameter estimation is essential for reliable electrical modelling of photovoltaic (PV) modules. This study proposes an ameliorated social group optimization algorithm (ASGO) to identify the five parameters of the single diode model and the seven parameters of the double diode model. The improving phase incorporates Gaussian mutation, an adaptive current-to-best/2 mutation strategy, and local search to enhance population diversity and reduce premature convergence, while retaining the original acquiring phase. Parameter identification minimizes the root mean square error between measured and calculated currents. Performance is evaluated against conventional social group optimization using the RTC France cell, the Photowatt PWP-201 module. Each algorithm is executed independently 50 times to assess accuracy, consistency, convergence, and computational time. Although both methods frequently achieve similar minimum errors, ASGO produces lower average errors and more consistent outcomes across repeated runs. Current–voltage curves reconstructed using parameters from the worst runs remain closer to experimental measurements with ASGO. Exploration and diversity analyses support its improved search behaviour. These results demonstrate that ASGO provides a reliable estimation approach for the tested photovoltaic models, with less controlling parameters tuning, faster convergence in iterations and a modest increase in execution time.
Keywords
