A Clustering Method of Generalized Load Characteristics Based on AP Clustering Algorithm

Abstract

The invention discloses a generalized load characteristic clustering method based on an AP clustering algorithm. The method comprises the following steps: collecting root bus node wind power and generalized load data, and calculating a field average velocity, a fluctuation velocity and fluctuation intensity; determining minimum time span and clustering time interval of the fluctuation intensity; segmenting all training sample data of generalized load node power according to the clustering time interval in a unified manner and constructing a cluster feature vector; endowing sample data with bias parameters; clustering on the basis of the AP clustering algorithm and by taking the feature vector as an index, and determining a clustering center; if the clustering center of a sample point meets the set clustering condition, outputting a clustering result, otherwise, returning to the former step until an accurate clustering result is obtained. By adopting the method provided by the invention, the problem that the conventional modeling method cannot describe node characteristic random variation is effectively solved.

Description

technical field [0001] The invention relates to a generalized load characteristic clustering method based on an AP clustering algorithm. Background technique [0002] Large-scale wind power grid integration has a great impact on the safe, economical and stable operation of the power system. The integration of wind power has changed the situation of pure power consumption by loads, making it possible for load nodes to send power back to the grid. Wind power has random volatility and intermittent nature, while load itself has time-varying nature. The interaction between the two aggravates the uncertainty of generalized load nodes, which will have a great impact on system power flow distribution, system stability, and safe operation of the power grid. , so the generalized load node characteristic modeling considering the random fluctuation of wind power and time-varying load is of great significance to the power system analysis. [0003] The traditional modeling method adopts ...

AI Technical Summary

Problems solved by technology

The traditional load modeling clustering method is effective in solving the time-varying load, but the clustering method used in it needs to artificially set the number of clusters, cluster centers, etc., which is relatively subjective and not universal when considering complex scenarios of wind power access. applicability

Due to the continuous expansion of the scale of wind power access and the gradual increase in penetration rate, the situation of simple power consumption of traditional load nodes has been changed, and the composition of load nodes and the direction of power flow have undergone essential changes. Traditional simple clustering strategies and clustering methods cannot adapt to new scenarios. Therefore, it is an urgent problem to study the reasonable clustering and synthesis of generalized load uncertainty under the new scenario of wind power integration.

Images