Clustering methodologies for wind power sampling in large scale network generating system reliability studies

Abstract EN

Given the highly fluctuating behavior of wind energy, statistical models must be developed to quantify the impact of this renewable energy on the electrical system reliability.

In this study, several wind sampling methodologies are introduced and compared in order to investigate the impact of wind production over large scale networks (and, more particularly, the Belgian network case).

In that context, dispersed generators are clustered on basis of wind statistical behavior and several correlation levels with the load are considered to investigate the worst case for the evaluation of wind impact on large scale generating system reliability.

Moreover, clustering hypothesis initially made for narrow generating systems are confronted to larger areas and compared with a more realistic hierarchical wind model inside each defined cluster.

In fact, the proposed intra cluster hierarchical methodology is based on the extraction of the correlation between all the statistical wind distributions established inside the cluster.

This last method consequently leads to a more realistic model in the case of large scale wind clusters as it will permit to distinguish correlation and independence between wind speeds inside the same cluster.

All the wind sampling methodologies are introduced in a Monte Carlo simulation and are compared thanks to a well-being analysis.

It is finally shown that the proposed technique will assist system planners and transmission system operators to qualitatively assess the system impact of wind production and to provide adequate input for the managerial decision process in presence of increased wind penetration.

Note that those methodologies are, here, applied to the Belgian network.