TY - JOUR
T1 - Intelligent detection of the pv faults based on artificial neural network and type 2 fuzzy systems
AU - Janarthanan, Ramadoss
AU - Maheshwari, R. Uma
AU - Shukla, Prashant Kumar
AU - Shukla, Piyush Kumar
AU - Mirjalili, Seyedali
AU - Kumar, Manoj
N1 - Funding Information:
Funding: This work was partially funded by the Center for Artificial Intelligence and Research, Chennai Institute of Technology, vide funding number is CIT/CAIR/2021/004.
Publisher Copyright:
© 2021 by the authors. Licensee MDPI, Basel, Switzerland.
PY - 2021/10/1
Y1 - 2021/10/1
N2 - The real‐time application research on the Fuzzy Logic Systems (FLSs) and Artificial Neural Networks (ANN) is vast and, in this paper, a technique for a photovoltaic failure analysis using the type 2 FLS and ANN is proposed. The method is proposed to build T2 FLS with a guaranteed value equal to or lower than T2 and ANN. Several explanations are conducted to illustrate the effectiveness of the methodologies. It is found that both the type 2 Fuzzy and ANN can be configured for productive actions in applications for a PV fault analysis, and choice is typically applied. The methods discussed in this paper lay the groundwork for developing FLSs and ANNs with durable characteristics that will be extremely useful in many functional applications. The result demonstrates that specific fault categories can be detected using the fault identification method, such as damaged PV modules and partial PV unit shades. The average detection performance is similar in both ANN and fuzzy techniques. In comparison, both systems evaluated show approximately the same performance during experiments. The architecture of the type 2 fuzzy logic system and ANN with radial basic function, including the roles of the output port and the rules for identifying the type of defect in the PV structure is slightly different.
AB - The real‐time application research on the Fuzzy Logic Systems (FLSs) and Artificial Neural Networks (ANN) is vast and, in this paper, a technique for a photovoltaic failure analysis using the type 2 FLS and ANN is proposed. The method is proposed to build T2 FLS with a guaranteed value equal to or lower than T2 and ANN. Several explanations are conducted to illustrate the effectiveness of the methodologies. It is found that both the type 2 Fuzzy and ANN can be configured for productive actions in applications for a PV fault analysis, and choice is typically applied. The methods discussed in this paper lay the groundwork for developing FLSs and ANNs with durable characteristics that will be extremely useful in many functional applications. The result demonstrates that specific fault categories can be detected using the fault identification method, such as damaged PV modules and partial PV unit shades. The average detection performance is similar in both ANN and fuzzy techniques. In comparison, both systems evaluated show approximately the same performance during experiments. The architecture of the type 2 fuzzy logic system and ANN with radial basic function, including the roles of the output port and the rules for identifying the type of defect in the PV structure is slightly different.
KW - Artificial neural network
KW - Machine learning
KW - Photovoltaic (PV) fault detection
KW - Type 2 fuzzy logic systems
UR - http://www.scopus.com/inward/record.url?scp=85117354577&partnerID=8YFLogxK
U2 - 10.3390/en14206584
DO - 10.3390/en14206584
M3 - Article
AN - SCOPUS:85117354577
SN - 1996-1073
VL - 14
JO - Energies
JF - Energies
IS - 20
M1 - 6584
ER -
