Abstract：In the new power system，the generation method and operating characteristics of new energy are

significantly different from traditional synchronous machines，and the safe and stable operation of frequency faces

new challenges and opportunities. When there is a large frequency fluctuation caused by high-power disturbances in

the power grid，in order to adjust the system frequency，it is necessary to manually adjust the output of the units in a timely manner. The rationality of power allocation of frequency modulation units is of great significance to energy

utilization efficiency and power grid stability. Based on the deep deterministic policy gradient（DDPG）algorithm

in deep reinforcement learning，a reward function was designed based on the frequency deviation and regional

control error of the regional system. Multiple scenarios were set to train the intelligent agent，and scene weights

were assigned according to the disturbance size. The weighted power allocation factor of the general policy set was

output，ultimately achieving a reasonable allocation of unit frequency modulation power. A joint simulation

platform was established based on Python calling PSS/E，and the 44 node IEEE standard power system model in the

Nordic region was used as a simulation example to verify the feasibility of using the joint simulation platform to

generate optimization strategy results.

        Key words：new power system；frequency stabilization；frequency modulation power distribution；depth

deterministic policy gradient（DDPG）algorithm

     Format Citation:李明，魏承志，郭小易,等. 基于DDPG的新型配电网调频功率分配仿真与控制研究[J].电气传动,2026,56(08):39-48. LI Ming，WEI Chengzhi，GUO Xiaoyi,et al. Research on simulation and control of frequency regulation power distribution in new distribution network based on DDPG [J].Electric Drive, 2026,56(08): 39-48