Discrete control model Q-learning for an energy storage system with a hydrogen unit of an autonomous hybrid power plant of a railway substation

The paper considers the prospects for creating autonomous hybrid power plants using renewable energy sources and hydrogen as energy storage systems, as well as storage batteries, for the railway power supply system. A comparative analysis of various energy storage systems is carried out. A simulatio...

Full description

Saved in:
Bibliographic Details
Published in:International journal of hydrogen energy 2024-12, Vol.93, p.704-714
Main Authors: Matrenin, Pavel, Ghulomzoda, Anvari, Safaraliev, Murodbek, Tavlintsev, Alexander
Format: Article
Language:English
Subjects:
Energy storage systems
Hydrogen
Hydrogen storage
Predictive control
Q-learning
Railway electrification system
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The paper considers the prospects for creating autonomous hybrid power plants using renewable energy sources and hydrogen as energy storage systems, as well as storage batteries, for the railway power supply system. A comparative analysis of various energy storage systems is carried out. A simulation model has been created that takes into account the characteristics of electric rolling stock trains, a model of the contact network of the inter-station zone of the railway stage, and a model of the energy storage system. Managing the flow of electricity in an autonomous hybrid power plant requires considering forecasts of electricity consumption and generation and the level of charge of the energy storage device. To solve the problem of synthesizing a control algorithm, the use of matrix Q-learning is proposed. The novelty of the study lies in the proposed approach of applying Q-learning to the problem under consideration based on discretization of input and output parameters of the model and verification of the approach on a simulation model built for a real railway section between the Yaya and Izhmorskaya stations (Kemerovo region of the Russian Federation). It is shown that the use of the proposed system makes it possible to equalize the voltage in the network between traction substations and provide a significant increase in the capacity of the railway section, and the proposed method of matrix Q-learning makes it possible to synthesize an effective algorithm for controlling the energy storage system as part of an autonomous hybrid power plant.
ISSN:0360-3199
DOI:10.1016/j.ijhydene.2024.10.397