A Novel Nonlinear Deep Reinforcement Learning Controller for DC–DC Power Buck Converters

Gheisarnejad, Meysam; Farsizadeh, Hamed; Khooban, Mohammad Hassan

doi:10.1109/tie.2020.3005071

Cited by 110 publications

(33 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…where T sw is periods of excitation waveform, dB dt is slope of flux density varying from time and ∆B is flux density variation within one switching period. k i is defined as (9).…”

Section: B Magnetic Loss Model 1) Inductor Design Methodmentioning

confidence: 99%

“…As for its application in power electronics, Tang proposed a deep reinforcement learning aided method to optimize the triple phase shift control scheme in order to achieve the high efficiency of a dual active bridge converter [8]. Gheisarnejad proposed a RL-aided controller for DC-DC power buck converters [9]. RL is implemented to reduce the observer estimation error.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Deep Reinforcement Learning for DC-DC Converter Parameters Optimization

Tian

Cobaleda

Martínez

2022

2022 IEEE 31st International Symposium on Industrial Electronics (ISIE)

View full text Add to dashboard Cite

Reinforcement learning (RL) is a type of machine learning in which an agent teaches itself by interacting with the environment. A RL-based parameter optimization method is proposed to improve the efficiency of a DC-DC power converter. More specifically, deep Q network (DQN) methods are utilized to optimize the power converter's parameter designs under current, voltage ripple, and volume limitations. Spice simulation is used to determine power losses on semiconductors. Combined with an optimal design of the inductor, the overall efficiency of power converters is obtained. The results show that an optimal design is obtained by using the DQN algorithm.

show abstract

“…where T sw is periods of excitation waveform, dB dt is slope of flux density varying from time and ∆B is flux density variation within one switching period. k i is defined as (9).…”

Section: B Magnetic Loss Model 1) Inductor Design Methodmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Deep Reinforcement Learning for DC-DC Converter Parameters Optimization

Tian

Cobaleda

Martínez

2022

2022 IEEE 31st International Symposium on Industrial Electronics (ISIE)

View full text Add to dashboard Cite

show abstract

“…Such method of combining deep NN with RL is called deep reinforcement learning (DRL) which has the advantages of overcoming the "dimensional curse", and does not need system identification steps that may be difficult to obtain in practice. Based on these advantages, the DRL-based methods have been applied to the optimization of wind power forecast uncertainty [28], multiscenario emergency controller [29], power electronic controller [30], and EV charging scheduling. Specially, [31] considers the randomness of commuting behavior and the uncertainty of electricity price, and the authors apply a naive datadriven deep Q network (DQN) algorithm to obtain a charging strategy without any model information.…”

Section: Introductionmentioning

confidence: 99%

Electric Vehicle Charging Management Based on Deep Reinforcement Learning

Cao

et al. 2022

Journal of Modern Power Systems and Clean Energy

View full text Add to dashboard Cite

A time-variable time-of-use electricity price can be used to reduce the charging costs for electric vehicle (EV) owners. Considering the uncertainty of price fluctuation and the randomness of EV owner' s commuting behavior, we propose a deep reinforcement learning based method for the minimization of individual EV charging cost. The charging problem is first formulated as a Markov decision process (MDP), which has unknown transition probability. A modified long short-term memory (LSTM) neural network is used as the representation layer to extract temporal features from the electricity price signal. The deep deterministic policy gradient (DDPG) algorithm, which has continuous action spaces, is used to solve the MDP. The proposed method can automatically adjust the charging strategy according to electricity price to reduce the charging cost of the EV owner. Several other methods to solve the charging problem are also implemented and quantitatively compared with the proposed method which can reduce the charging cost up to 70.2% compared with other benchmark methods.

show abstract

“…However, the transient behavior of the converter has not been considered and the converter is not well controlled in the severe load and input disturbances. Deep reinforcement learning based controlling technique has been recently employed in several applications [36]- [38]. These researches apply the reinforcement learning on the different strategies such as sliding mode and model predictive for enhancing the transient and steady state behavior of the DC-DC converters.…”

Section: Iintroductionmentioning

confidence: 99%

Application of Mixture of Experts in Machine Learning-Based Controlling of DC-DC Power Electronics Converter

et al. 2022

View full text Add to dashboard Cite

Regardless of the application in which power electronic converters are deployed, their desired performances crucially depend on the controlling strategy while different impressive parameters are varied. This paper offers a novel controlling strategy originated from the mixture of two well-known controlling techniques, namely feedback (FBC) and model predictive (MPC) controllers. It uses the advantages of the above-mentioned controllers while their drawbacks or limitations are covered by each other using the mixture of experts (MoE) technique. Two neural networks for capturing the features of MPC and FBC along with a gating network as the main tool of MoE are employed in order to optimize the controlling of the DC-DC power electronic converters. These networks are trained through a set of pair data as the input vector and the target data. The results reveal that better performance can be obtained via benefit exploitation of both controlling techniques using a comprehensive MoE. The dynamic and steady state errors are decreased by 5% and 8%, respectively which demonstrate a global enhancement in the controlling of the DC-DC power electronic converters.

show abstract

A Novel Nonlinear Deep Reinforcement Learning Controller for DC–DC Power Buck Converters

Cited by 110 publications

References 45 publications

Deep Reinforcement Learning for DC-DC Converter Parameters Optimization

Deep Reinforcement Learning for DC-DC Converter Parameters Optimization

Electric Vehicle Charging Management Based on Deep Reinforcement Learning

Application of Mixture of Experts in Machine Learning-Based Controlling of DC-DC Power Electronics Converter

Contact Info

Product

Resources

About