Placement and sizing of distributed generation in distribution networks considering energy storage scheduling optimization

Abstract

Stabilizing the power flow distribution in distribution networks and determining the connection locations and capacities of distributed generation are crucial issues in optimizing the operation of distribution networks with distributed generation. This paper proposes an energy storage scheduling and optimization model based on deep reinforcement learning （deep RL） to match the relationship between distributed energy resource allocation and electricity load demand， thereby stabilizing power flow distribution in distribution networks with high penetration rates. Using line losses and voltage fluctuations as the loss functions， the paper proposes a decision-making model for placement and sizing of distributed generation based on multi-objective genetic algorithm. Testing is conducted on the IEEE 14-bus system， and the results indicate that the algorithm can effectively select the optimal connection locations and capacities for distributed generation， reducing overall line losses while ensuring voltage amplitude remains stable.

Keywords

• distributed generation
• deep rl
• energy storage optimization
• multi-objective genetic algorithm
• placement and sizing