Research on a data-driven evaluation model for assessing the capacity of distribution grids to accommodate distributed photovoltaic connections

Key Points

• The aim is to develop a model for assessing how well distribution grids can manage connections from distributed photovoltaic sources.
• Developed a hosting capacity assessment model using deep learning and robust chance-constrained optimization.
• Conducted case studies to validate the model's effectiveness and accuracy under varying PV penetration levels.
• Analyzed the influence of photovoltaic output uncertainty on the performance of AC/DC hybrid distribution networks.
• The proposed model significantly enhances the efficiency of hosting capacity assessments while retaining accuracy.
• It effectively demonstrates the effects of PV output uncertainty on hybrid distribution networks' hosting capacity.
• Results indicate improved operational security and better integration of distributed PV.

Abstract