Lightweight actor-critic generative adversarial networks for real-time smart generation control of microgrids

Han, Kunlun; Yang, Kai; Yin, Linfei

doi:10.1016/j.apenergy.2022.119163

Cited by 20 publications

(8 citation statements)

References 52 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Although optimal scheduling problems using DQN can solve the problem of highdimensional observation space, the method based on DQN can only deal with discrete action spaces. To solve dispatch problems with continuous-valued state and action variables, many scholars have begun to focus on DRL algorithms based on policy gradient, including deep deterministic policy gradient (DDPG) [50], actor-critic (AC) [47] and its variants, such as advantage actor-critic (A2C) [48], asynchronous advantage actor-critic (A3C) [49], and soft actor-critic (SAC) [56]. Nevertheless, parameter updates should be avoided to improve the stability of the DRL-agent training.…”

Section: Deep Reinforcement Learning Methodsmentioning

confidence: 99%

Low‐carbon economic dispatch of the combined heat and power‐virtual power plants: A improved deep reinforcement learning‐based approach

Tan

Shen

et al. 2022

IET Renewable Power Gen

View full text Add to dashboard Cite

To realize the national strategies for carbon emission reduction, electric power industries should undergo reforms to cope with the multiple challenges of decarbonization, marketization, and energy transition. How to design a dispatch strategy that considers both low‐carbon demand and economic cost has become a major concern in integrated energy systems. To realize multi‐energy complementation and low‐carbonization, a scheduling framework of combined heat and power‐virtual power plants (CHP‐VPPs) considering carbon capture and power‐to‐gas is proposed. The stochastic dynamic optimization problem is modelled as a Markov decision process. An improved deep reinforcement learning‐based algorithm named multi‐level backtracking prioritized experience replay‐twin delayed deep deterministic policy gradient (MBEPR‐TD3) is employed to solve the low‐carbon economic dispatch problem. The results show that: (1) The profits have been increased by 85.8% and carbon emissions have been reduced by 30.3% because of the addition of power‐to‐gas in CHP‐VPP; (2) the revenues have been improved by 22.24% and carbon emissions have been decreased by 37.04% owing to the introduction of carbon capture and trading; (3) compared with results using DQN, DDPG, and TD3, the optimal dispatch strategy obtained by MBEPR‐TD3 has higher reward returns and more stable convergence.

show abstract

Section: Deep Reinforcement Learning Methodsmentioning

confidence: 99%

Low‐carbon economic dispatch of the combined heat and power‐virtual power plants: A improved deep reinforcement learning‐based approach

Tan

Shen

et al. 2022

IET Renewable Power Gen

View full text Add to dashboard Cite

show abstract

“…It is a technique based on a program that learns from data and that works similarly to the way the human brain works. The common neuronal networks are feedforward neural networks (FFNNs), 17 convolutional neural networks (CNNs), 18 recurrent neural networks (RNNs), 19 deep belief networks (DBNs), 20 generative adversarial networks (GANs), 21 and spiking neural networks (SNNs). 22 This study uses a back propagation (BP) neural network, which is one of the FFNNs.…”

Section: Composition Of Neural Network Prediction Modelmentioning

confidence: 99%

A neural network algorithm and its prediction model towards the full color phase mixing process of colored fibers

Sun

Liu²,

Wang³

et al. 2022

Textile Research Journal

View full text Add to dashboard Cite

Aiming at the demand of color matching techniques in the spinning process, a neural network prediction model is constructed in this research study, and the gridded full color phase mixing space of colored fibers is used as the sample space. Subsequently, 30 grid points are employed as training samples, while another 30 grid points are adopted as testing samples, in which the parameters of the input, hidden, and output layers are optimized. Additionally, the neural network prediction model is constructed by training samples, and validated by testing samples. Lastly, a neural network prediction model is applied to implement the prediction of color and mixing ratios for any point within the full color phase mixing model. Through the assessment of the testing samples, the predicted results for the colors of the grid point samples showed an average color difference of 1.29 (minimum was 0.22 and maximum was 2.97); the forecasts for the mixing ratios of the colored fibers were that the range of the mean absolute error for the mixing ratios of individual samples was from 0.01% to 0.18%, and the mean absolute error for the mixing ratios of all samples was 0.21%. The experimental results indicated that the proposed neural network model has a relatively advanced prediction accuracy.

show abstract

“…Currently, a data-driven approach cannot avoid involving deep learning methods [9,10]. Deep learning can be classifed as convolutional neural networks [8], deep neural networks [11], deep reinforcement learning [12], and deep forest algorithms [13]. Deep learning can in turn be classifed as classifcation algorithms, prediction algorithms, and control algorithms [14].…”

Section: Introductionmentioning

confidence: 99%

“…Recently, numerous deep reinforcement technologies have been combined to achieve better control performances in more complex scenarios. For example, traditional controllers + deep reinforcement learning [4], modal decomposition + generative adversarial networks [12], and twindelayed DDPG + DDPG [20] are combined to address the frequency control problems of novel power systems; Markov chain and isoprobabilistic transformation are combined for capacitor planning [21]. Overall, the primary contributions of this work are summarized as follows:…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Multilayer Deep Deterministic Policy Gradient for Static Safety and Stability Analysis of Novel Power Systems

Long

Zhao

et al. 2023

International Transactions on Electrical Energy Systems

View full text Add to dashboard Cite

More and more renewable energy sources are integrated into novel power systems. The randomness and fluctuation of such renewable energy sources bring challenges to the static stability and safety analysis of novel power systems. In this work, a multilayer deep deterministic policy gradient is proposed to address the fluctuation of renewable energy sources. The proposed method is stacked with multilayer deep reinforcement learning methods that can be continuously updated online. The proposed multilayer deep deterministic policy gradient is compared with other deep learning algorithms. The feasibility, effectiveness, and superiority of the proposed method are verified by numerical simulations.

show abstract

Lightweight actor-critic generative adversarial networks for real-time smart generation control of microgrids

Cited by 20 publications

References 52 publications

Low‐carbon economic dispatch of the combined heat and power‐virtual power plants: A improved deep reinforcement learning‐based approach

Low‐carbon economic dispatch of the combined heat and power‐virtual power plants: A improved deep reinforcement learning‐based approach

A neural network algorithm and its prediction model towards the full color phase mixing process of colored fibers

Multilayer Deep Deterministic Policy Gradient for Static Safety and Stability Analysis of Novel Power Systems

Contact Info

Product

Resources

About