This paper presents a method based on an improved sparrow search algorithm (ISSA) and optimized random forest (RF) model
aiming to improve the accuracy of fault diagnosis in photovoltaic arrays. Firstly
the photovoltaic fault data set is constructed by building a photovoltaic array to simulate five fault conditions and extracting fault vectors. Secondly
the optimization comparison among sparrow search algorithm (ISSA)
grey wolf optimization (GWO)
particle swarm optimization (PSO) and ISSA by test function is performed. The results indicate that ISSA is better than other algorithms in terms of mean value and standard deviation
showing better robustness. Then
the performance of ISSA-RF is analyzed using the simulation data set of photovoltaic array fault. It can be found that the overall accuracy of ISSA-RF reaches 97.06%
which is 6.94 percentage points higher than the traditional RF model. Finally
the data set under five different working conditions including open circuit
short circuit
shade
aging and normal operation are collected for testing ISSA-RF. The research outcomes prove that this model has high classification efficiency and accuracy
CHEN S Q, YANG G J, GAO W, et al.Photovoltaic fault diagnosis via semisupervised ladder network with string voltage and current measures[J]. IEEE journal of photovoltaics, 2021, 11(1): 219-231.
ANDERSON A J.Photovoltaic translation equations: a new approach. Final subcontract report[R]. Colorado, National Renewable Energy Laboratory, 1996.
GHAREHCHOPOGH F S, NAMAZI M, EBRAHIMI L, et al.Advances in sparrow search algorithm: a comprehensive survey[J]. Archives of computational methods in engineering, 2023, 30(1): 427-455.
TIZHOOSH H R.Opposition-based learning: a new scheme for machine intelligence[C]//International Conference on Computational Intelligence for Modelling, Control and Automation and International Conference on Intelligent Agents, Web Technologies and Internet Commerce (CIMCA-IAWTIC'06). Vienna, Austria, 2006: 695-701.
HAFEZ A Z, YOUSEF A M, HARAG N M.Solar tracking systems: technologies and trackers drive types:areview[J]. Renewable and sustainable energy reviews, 2018, 91: 754-782.