Reinforcement learning for electricity dispatch in grids with high intermittent generation and energy storage systems: A case study for the Brazilian grid

Pinheiro, Vinícius de Carvalho Neiva; Francato, Alberto Luiz; Powell, Warren B.

doi:10.1002/er.5551

Cited by 11 publications

(5 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…13 There is still room for enhancing conventional timeseries and ML techniques in terms of accuracy and robustness. 14 To bridge the knowledge gap with scientific originality, linear and nonlinear methods were employed with metaheuristic optimization algorithms to establish multiple-step ahead power consumption prediction models, which can be utilized to estimate the future trends of energy consumption and assist decision-makers in formulating regulations regarding energy demands and utilization. Its predictions support power company to forecast energy consumption to efficiently dispatch regional energy capacity and keep the electricity supply and demand in balance for residential buildings in sustainable cities.…”

Section: Introductionmentioning

confidence: 99%

Multistep energy consumption forecasting by metaheuristic optimization of time‐series analysis and machine learning

Chou

Truong

2020

Int J Energy Res

View full text Add to dashboard Cite

Summary This study aims to develop a novel forecasting system that optimizes linear time‐series with nonlinear machine learning models to identify the historical pattern of regional energy consumption. The linear time‐series model, Seasonal AutoRegressive Integrated Moving Average (SARIMA), was applied to simulate the linear component, while the least squares support vector regression (LSSVR) model was used to capture the nonlinear component of time series energy data and combine the linear and nonlinear components. Several optimization algorithms were investigated using high‐dimension mathematical benchmark functions to compare their outcomes for applying in the proposed forecasting system. Wilcoxon rank‐sum test and convergence graphs for Jellyfish Search (JS) and other parameter‐less algorithms (ie, Teaching‐Learning‐Based Optimization (TLBO) and Symbiotic Organisms Search, SOS algorithms) indicate that the JS optimizer finds the best global solutions in mathematical tests. Then, three real energy time‐series datasets from a regional transmission organization that coordinates the movement of wholesale electricity were used to evaluate various forecasting methods. The analytical results confirm that the proposed system, JS‐LSSVR(SARIMA, LSSVR), can predict multi‐step ahead time series energy consumption with higher accuracy than the linear model (ie, SARIMA), nonlinear model (ie, LSSVR), hybrid model (ie, JS‐LSSVR), hybrid systems (ie, TLBO‐LSSVR(SARIMA, LSSVR) and TLBO‐LSSVR(SARIMA, LSSVR)), and prior studies. Numerical experiments show that the JS‐LSSVR(SARIMA, LSSVR) system can forecast energy consumption 1 week ahead efficiently (from 9.8 to 21.4 seconds on average). Notably, the proposed technique only requires 4 inputs vs 7 to 32 inputs of other models in the literature. Thus, a power company can apply the proposed system to forecast energy consumption to efficiently dispatch regional energy capacity and keep the electricity supply and demand in balance for residential buildings in sustainable cities.

show abstract

Section: Introductionmentioning

confidence: 99%

Multistep energy consumption forecasting by metaheuristic optimization of time‐series analysis and machine learning

Chou

Truong

2020

Int J Energy Res

View full text Add to dashboard Cite

show abstract

“…Reinforcement learning (RL) does not require external training data such as historical input-output data, which would make this approach in favor of the dispatch optimization problem. It builds its knowledge by an iterative approach which is similar to how the human gains his experience, which means that the model generates its training data and trains itself [ 41 ]. This is achieved by using a system of reward and penalty for the model when it performs correctly or incorrectly, respectively.…”

Section: Methodsmentioning

confidence: 99%

“…Constraints of the model of Hua, Qin [ 28 ] are related to power only, the current research advances this by addressing the interconnecting heating, cooling, and power constraints and considering a CCHP with technical efficiency modeling in the problem. Other studies introduced reinforcement learning in the dispatch of power plants but mainly to improve the forecasting of uncertain parameters such as solar power, wind power, or market prices [ 29 , 30 ].…”

Section: Literature Reviewmentioning

confidence: 99%

Assessing and comparing a DDPG model and GA optimization for a heat and power virtual power plant operating in a power purchase agreement scheme

Elgamal,

Shahrestani,

Vahdati

2024

Heliyon

View full text Add to dashboard Cite

“…These techniques enable decision-makers to make modifications based on power requirements and usage by calculating the progressive nature of power usage [13]. Energy storage and highintermittent generating networks use reinforcement learning to suggest using deterministic look-ahead rule-based changes for arrangement and a straightforward Policy Function Approximation (PFA) for the real-time operation to create rules for energy dispatch with undefined prognoses [14]. An overview of real-time applications of machine learning for predicting growth trends in various energy systems uses three well-known forecasting engines to conduct a thorough review of supervised-based machine learning methods.…”

Section: Introductionmentioning

confidence: 99%

Energy Forecasting Information System of Optimal Electricity Generation using Fuzzy-based RERNN with GPC

2023

KSII TIIS

View full text Add to dashboard Cite

In this paper, a hybrid fuzzy-based method is suggested for determining India's best system for power generation. This suggested approach was created using a fuzzy-based combination of the Giza Pyramids Construction (GPC) and Recalling-Enhanced Recurrent Neural Network (RERNN). GPC is a meta-heuristic algorithm that deals with solutions for many groups of problems, whereas RERNN has selective memory properties. The evaluation of the current load requirements and production profile information system is the main objective of the suggested method. The Central Electricity Authority database, the Indian National Load Dispatch Centre, regional load dispatching centers, and annual reports of India were some of the sources used to compile the data regarding profiles of electricity loads, capacity factors, power plant generation, and transmission limits. The RERNN approach makes advantage of the ability to analyze the ideal power generation from energy data, however the optimization of RERNN factor necessitates the employment of a GPC technique. The proposed method was tested using MATLAB, and the findings indicate that it is effective in terms of accuracy, feasibility, and computing efficiency. The suggested hybrid system outperformed conventional models, achieving the top result of 93% accuracy with a shorter computation time of 6814 seconds.

show abstract

Reinforcement learning for electricity dispatch in grids with high intermittent generation and energy storage systems: A case study for the Brazilian grid

Cited by 11 publications

References 32 publications

Multistep energy consumption forecasting by metaheuristic optimization of time‐series analysis and machine learning

Multistep energy consumption forecasting by metaheuristic optimization of time‐series analysis and machine learning

Assessing and comparing a DDPG model and GA optimization for a heat and power virtual power plant operating in a power purchase agreement scheme

Energy Forecasting Information System of Optimal Electricity Generation using Fuzzy-based RERNN with GPC

Contact Info

Product

Resources

About