A data-driven time-delay compensation strategy for ancillary service of the distribution photovoltaic generation system

Abstract

With the increasing integration of distributed photovoltaic (PV) generation, the distribution network has met many challenges in operation and control. Ancillary services on PV generation systems become a necessary function to enhance the distribution network operating stability and resilience. However, due to the complexity of the control framework and the outer data dependency, the time delay on the measured data may bring a significant influence on the effectiveness of the PV system ancillary service. To solve the time delay influence, in this paper, a data-driven time-delay compensation strategy via the long short-term memory (LSTM) method is proposed. The proposed compensation strategy could realize the measured data compensation caused by the communication or calculation delay to maintain the accuracy of measured data that is input into the PV system ancillary service. Besides the LSTM-based method, the data-driven time-delay compensation strategy also includes a LSTM activation control to realize the smooth activation of the compensation strategy into the PV generation system. A modified IEEE 123 bus system with multiple distributed PV generation systems integration is conducted to verify the performance of the proposed compensation strategy. The simulation results indicate that the proposed data-driven time-delay compensation strategy could significantly improve the frequency performance of the PV ancillary service. In addition, the simulation results also show that the LSTM has a strong generalization ability for delay time constant and can deal with the random time delay caused by communication and disturbances in distribution systems.