Coordinated management of aggregated electric vehicles and thermostatically controlled loads in hierarchical energy systems

This paper presents an energy management model for electric vehicles (EVs) and thermostatically controlled loads (TCLs) in intelligent energy systems based on the transactive control of aggregators. The management strategy will penetrate through three physical layers of electricity networks: transmission layer, distribution layer, and behind-meter layer. In the proposed framework, the aggregated EVs are modeled as a battery energy storage system (BESS), and the aggregated TCLs are modeled as a virtual energy storage system (VESS) at the behind-meter layer. A deep learning method, namely a hybrid of convolutional neural networks and long shortterm memory (CNN-LSTM), is used to forecast the local loads of EVs and TCLs. The aggregators can dispatch these controllable loads directly as demand management to fit the predicted load curve. Peer-to-peer (P2P) trading is realized at the distribution level, and distributed optimization is utilized since the information between each aggregator is opaque. The primal problem is decoupled into subproblems of aggregators. The sub-gradient method is employed to update the multipliers of each decomposed Lagrange function. After the local energy transaction is cleared at the distribution level, wind generators and thermal generators are centrally dispatched at the transmission level based on the conventional optimal power flow model. The proposed hierarchy framework is verified in the IEEE 30-bus system. Simulation results reveal that the scalability issue of single-layer centralized dispatch can be well addressed, and end-users' information privacy can be protected. The coordinated management of EVs and TCLs also brings in economic and environmental benefits.

Automated summary

This paper introduces an energy management model for electric vehicles and thermostatically controlled loads in intelligent energy systems, utilizing a hierarchical management strategy to address energy demand management. It involves distributed optimization and peer-to-peer trading, improving end-users' information privacy protection and providing economic and environmental benefits.