An Equivalence Modeling Method of Photovoltaic Power Station Based on Improved Fuzzy Clustering Algorithm

Abstract

The invention discloses an equivalent modeling method for a photovoltaic power station based on an improved fuzzy clustering algorithm. The steps include: 1. determining the clustering index of each photovoltaic power generation unit and performing standardized processing; 2. setting clustering parameters, and initializing a membership degree matrix; 3. Clustering and grouping photovoltaic power generation units by improving fuzzy clustering algorithm; 4. According to the clustering results, calculate equivalent parameters and merge photovoltaic power generation units; 5. Verify the correctness of the equivalent model through error analysis. The invention can carry out scientific and effective grouping and equivalent for many photovoltaic power generation units, thereby improving the accuracy of the equivalent model of large-scale photovoltaic power station.

Description

technical field [0001] The invention relates to the field of power system simulation analysis, in particular to an equivalent modeling method for a large-scale grid-connected photovoltaic power station. Background technique [0002] With the gradual depletion of global fossil energy, photovoltaic power generation has developed rapidly in countries all over the world due to its clean, non-polluting, renewable, and huge reserves. In recent years, photovoltaic grid-connected simulation analysis has become a research hotspot in the field of power systems. The traditional simulation method establishes detailed simulation models for all grid-connected photovoltaic power generation units to achieve accurate simulation of the transient steady-state performance of photovoltaic grid-connected. However, with the continuous expansion of the scale of photovoltaic power plants, for a large number of centralized photovoltaic grid-connected units, the traditional modeling method not only c...

AI Technical Summary

Problems solved by technology

In FCM, the sum of the membership degrees of the same sample belonging to all classes is 1, which makes it very sensitive to noise and outlier points; the iterative descent algorithm adopted by FCM is sensitive to the initialized cluster center or membership matrix, Can not guarantee to converge to the global optimal solution, may converge to local extremum or saddle point

Images