The distribution network reconfiguration (DNR) aims at minimizing the power losses and improving the voltage profile. Traditional model-based methods exactly need the network parameters to derive the optimal configuration of the distribution network. This paper proposes a DNR method based on model-free reinforcement learning (RL) approach. The proposed method adopts NoisyNet deep Q-learning network (DQN), by which the exploration can be automatically realized without need of tuning the exploration parameters, in order to accelerate the training process and improve the optimization performance. The proposed method is validated by the simulation results.

Very-short-term load forecasting (VSTLF) predicts the load from minutes to 1-hour timescale. Effective forecasting is important for in-day scheduling of the power systems. In this paper, a VSTLF method based on empirical mode decomposition and deep neural network is proposed. The extreme point span is used to determine a proper empirical modal number, so as to successfully decompose the load data into different timescales, based on which the deep-neural-network-based forecasting model is established. The accuracy of the proposed method is verified by the testing results in this paper.

No abstract

It can be foreseen that the modern electric power system tends to have an increasing penetration of renewable energy sources (RESs) and controllable loads under ancillary services programmes, which will cause the power system inertia reduction. For the thermostatically controlled loads (TCLs) (e.g. air conditioners) have large potentials for providing ancillary services for power systems and are more flexible and cost-efficient than traditional energy storage system, in this study, an approach of modelling power system virtual inertia (VI) by TCLs is proposed, while maintaining customers' comfort level. Considering the TCLs' heterogeneous characteristic and response uncertainty, a load tracking control strategy based on proportional-integral controller is proposed to preserve the performance of VI control strategy. Besides, the coefficient γ is put forward for the first time to quantify and analyse the conversion efficiency between kinetic energy and thermal energy. Then, the method to design k vi based on the energy analysis is presented. Finally, some simulation results are provided to verify the effectiveness of the control scheme.