Research On Control Strategy And Transient Stability Assessment Method Of AC/DC Microgrid

Establishing a system with distributed generations(DGs)is an important development direction of modern power grids by using renewable energy to generate electricity.With the increase of penetration of DGs,AC/DC hybrid microgrids have emerged.Reasonable control strategy and accurate transient stability assessment are the basic requirements for ensuring the safe and stable operation of AC/DC hybrid micorgird.Firstly,the transient model of AC/DC hybrid microgrid system is established.Models of wind power generation systems and photovoltaic power generation systems are established,wherein the wind power generation system includes a constant frequency/constant speed wind power generation system and a constant frequency/variable speed wind power generation system.Moreover,the transient model of the energy storage system is established.Then the multi-time-scale based sub-area AC/DC hybrid microgrid control strategy is proposed.The whole control system is divided into three layers.The first layer is used to adjust the common DC bus load,and the time scale is 10ms.The second layer is applied to adjust the AC subsystem and the DGs in the DC subsystem,the time scale is 1s.The third layer is the upper layer control strategy,that is,the optimal coordination control,the time scale is 10s.The purpose of the first and the second layer control strategy is to ensure the stability of the active power in the AC and DC subsystems.Thus,the system is stable without fluctuations of the DGs.The purpose of the third layer control strategy is to achieve stable system level power fluctuations between subsystems through appropriate power exchange.When the DGs fluctuates,both the AC subsystem and the DC subsystem need to coordinate the shared power fluctuations.Therefore,the subsystem can coordinate with each other to improve the overall performance of the system.Finally,a transient stability assessment method for AC/DC hybrid microgrid based on convolutional neural network is proposed.The samples obtained from the system after the disturbance are used as input samples,and CNN is used to automatically extract features from the original data.The loss function sensitive to cost loss is proposed,which reduces the false detection of instability.The simulation results show that CNN can automatically extract features from massive data and improve the prediction accuracy of the system state.The sensitive loss function effectively improves the accuracy of predicting unstable samples.