Heterogeneous Aggregation Control Model for Hybrid EV Clusters with Random Charging Behavior

With a large population of electric vehicles (EVs) widely connecting to the grid, uniformly aggregating EVs as a system and controlling the system to provide considerable power transfer capacity is significant to support the supply-demand balance. The system includes hybrid EV clusters of different parameters with random arrival and departure, resulting in the problems of parameter heterogeneity and randomness in aggregation and control. The parameter heterogeneity incurs different charging response characteristics. So, it's of great difficulty to aggregate and control heterogeneous EVs. Besides, the aggregation accuracy is also affected by random arrival and departure numbers. To address these issues, an aggregation control model is proposed to integrate hybrid EV clusters uniformly. Firstly, an equivalent aggregation model for heterogeneous EVs is constructed and the radial basis function (RBF) neural network is used to obtain the equivalent aggregation parameters. Secondly, the modification of equivalent aggregation parameters due to random numbers is proposed to ensure the model's accuracy. Finally, the sliding mode control algorithm is applied to control the EV system's aggregated power for target power tracking services. The wind power and photovoltaic power as the consumption target are tracked by heterogeneous EVs of two scenarios with different heterogeneity in the simulation. The main verification of the proposed aggregation control performance is implemented. Besides, the special verification under more difference in parameter heterogeneity is also given. The metrics such as aggregation control accuracy, stability, user comfort are used to validate the model performance. Experiment results demonstrate that the proposed model can aggregate plentiful heterogeneous and random EVs as a uniform system for tracking the renewable energy output accurately.