Aiming at the equivalent modeling of concentrated PV power station and the traditional clustering algorithm's over-reliance on the initial clustering centers and the number of clusters
this paper proposes a clustering method for PV power station based on the affinity propagation (AP) algorithm
taking PV arrays with different solar irradiance as the object of study. The AP algorithm adopts the existing data points as the final clustering center
without specifying the initial clustering center and the number of clusters. Considering that the clustering index of PV power units needs to reflect the power output characteristics of PV power units
for this reason
the solar irradiance and the equivalent line impedance are taken as clustering indexes
and then the clustering algorithm is used to cluster the PV power units with different the solar irradiance and the equivalent line impedance into different categories
and the PV power units with similar categories are aggregated. At the same time
equate the parameters of the line impedance and inverters of the PV power station to establish a multi-machine equivalent model. Based on the PSCAD simulation platform
the clustering model of a typical 10 kV centralized PV power station is constructed by using the method of this paper
and the simulation results of the detail model
the AP clustering model
the C-FCM clustering model and the single-machine equivalent model are compared
which verifies the effectiveness of the clustering method of this paper under different solar irradiance of PV arrays.
HAN P P, LIN Z H, ZHANG J J, et al.Equivalent modeling of photovoltaic power plant based on factor analysis and correlation clustering[J]. IEEE access, 2019, 7: 56935-56946.
LI P X, GU W, LONG H, et al.High-precision dynamic modeling of two-staged photovoltaic power station clusters[J]. IEEE transactions on power systems, 2019, 34(6): 4393-4407.