Load and Photovoltaic Power Scenario Extraction Based on Copula Theory and Deep Convolutional Embedded Clustering

The inherent fluctuation of renewable energy sources, especially photovoltaic systems, challenges the planning and operation of distribution power systems. Extracting typical scenarios to give a reasonable description of load power and distributed photovoltaic power is a common method to address these challenges for renewable energy planning. Extracting typical scenarios includes two parts, namely scenario generation and scenario reduction. In this paper, a joint distribution model of load and photovoltaic power is first established based on the Copula theory. Then, combined with the correlation evaluation indexes, the optimal Copula function is selected as the joint probability distribution of load and photovoltaic power to generate multiple scenarios. Finally, based on the two-dimensional deep convolutional embedded clustering method, the typical matching load and photovoltaic power scenarios are obtained. Case analysis indicates that the typical load and photovoltaic power scenarios yielded by the proposed method align more closely with actual power curves, underscoring the effectiveness of the proposed method.