TY - JOUR
T1 - Maximum Power Point Estimation for Photovoltaic Strings Subjected to Partial Shading Scenarios
AU - Ma, Jieming
AU - Jiang, Haochuan
AU - Bi, Ziqiang
AU - Huang, Kaizhu
AU - Li, Xingshuo
AU - Wen, Huiqing
N1 - Publisher Copyright:
© 1972-2012 IEEE.
PY - 2019/3/1
Y1 - 2019/3/1
N2 - Partial shading is an unavoidable complication in the field of photovoltaic (PV) generation. Bypass diodes have become a standard feature of solar cell arrays to improve array performance under partial shading scenarios (PSS). However, the current-voltage and power-voltage characteristic data vary with different shading patterns. In this paper, a shading pattern detection algorithm is first proposed to estimate the number of shaded modules in a PV string. The result is then fed to a field-support vector regression (F-SVR) model, in which the measured features are transferred into a style-free high-dimension space prior to data training. The process of style-normalized transformation enables the features to be independent and identically distributed. Both simulations and experiments are conducted to evaluate the F-SVR model's ability to estimate the voltage at maximum power points. The results show that the proposed maximum power point estimation method can evidently reduce prediction errors.
AB - Partial shading is an unavoidable complication in the field of photovoltaic (PV) generation. Bypass diodes have become a standard feature of solar cell arrays to improve array performance under partial shading scenarios (PSS). However, the current-voltage and power-voltage characteristic data vary with different shading patterns. In this paper, a shading pattern detection algorithm is first proposed to estimate the number of shaded modules in a PV string. The result is then fed to a field-support vector regression (F-SVR) model, in which the measured features are transferred into a style-free high-dimension space prior to data training. The process of style-normalized transformation enables the features to be independent and identically distributed. Both simulations and experiments are conducted to evaluate the F-SVR model's ability to estimate the voltage at maximum power points. The results show that the proposed maximum power point estimation method can evidently reduce prediction errors.
KW - Field support vector regression (F-SVR)
KW - PV power generation system
KW - maximum power point (MPP) estimation
KW - partial shading scenarios (PSS)
KW - photovoltaic (PV) cells
UR - http://www.scopus.com/inward/record.url?scp=85057195360&partnerID=8YFLogxK
U2 - 10.1109/TIA.2018.2882482
DO - 10.1109/TIA.2018.2882482
M3 - Article
AN - SCOPUS:85057195360
SN - 0093-9994
VL - 55
SP - 1890
EP - 1902
JO - IEEE Transactions on Industry Applications
JF - IEEE Transactions on Industry Applications
IS - 2
M1 - 8540929
ER -