Dynamic equivalent modelling for active distributed network considering adjustable loads charging characteristics
Jingwen Wang,Jiehui Zheng,Zhigang Li,Qing‐Hua Wu
DOI: https://doi.org/10.1049/gtd2.13344
2024-12-11
IET Generation Transmission & Distribution
Abstract:For active distribution networks with a large number of renewable energy sources and adjustable loads connected to it, an equivalent modelling structure that is highly interpretable is proposed here. This paper transforms the problem of identifying key parameters of an equivalent model into a Markov decision process and solve for the key parameters by a four‐layer, three‐stage hierarchical reinforcement learning approach. As more renewable energy generators and adjustable loads such as electric vehicles are being connected to the power grids, load modelling of the distribution network becomes more complicated. Therefore, this paper explores a dynamic equivalent modelling method for active distribution network that takes into account electric vehicle charging. First of all the combination of integrated ZIP loads and motors is adopted as an equivalent model for active distribution networks. Subsequently, a four‐layer, tri‐stage deep reinforcement learning approach is used to solve the relevant key parameters of the proposed equivalent model. The method proposed in this paper fully utilizes the superiority of reinforcement learning in decision making, while the method combines the excellent feature extraction capability of deep learning. The method utilizes measurements obtained at boundary nodes to obtain an active distributed network equivalent model after a series of calculations. At the same time, adjustable loads are identified in detail. On the other hand, this method introduces a prioritized empirical playback mechanism, log‐cosh loss function, and adaptive operator to improve the computational efficiency of the method. From the simulation results, the present method is effective.
engineering, electrical & electronic