Centralized Energy Management Research Articles

Unveiling Genetic Reinforcement Learning (GRLA) and Hybrid Attention-Enhanced Gated Recurrent Unit with Random Forest (HAGRU-RF) for Energy-Efficient Containerized Data Centers Empowered by Solar Energy and AI

The adoption of renewable energy sources has seen a significant rise in recent years across various industrial sectors, with solar energy standing out due to its eco-friendly characteristics. This shift from conventional fossil fuels to solar power is particularly noteworthy in energy-intensive environments such as cloud data centers. These centers, which operate continuously to support active servers via virtual instances, present a critical opportunity for the integration of sustainable energy solutions. In this study, we introduce two innovative approaches that substantially advance data center energy management. Firstly, we introduce the Genetic Reinforcement Learning Algorithm (GRLA) for energy-efficient container placement, representing a pioneering approach in data center management. Secondly, we propose the Hybrid Attention-enhanced GRU with Random Forest (HAGRU-RF) model for accurate solar energy prediction. This model combines GRU neural networks with Random Forest algorithms to forecast solar energy production reliably. Our primary focus is to evaluate the feasibility of solar energy in meeting the energy demands of cloud data centers that utilize containerization for virtualization, thereby promoting green cloud computing. Leveraging a robust German photovoltaic energy dataset, our study demonstrates the effectiveness and adaptability of these techniques across diverse environmental contexts. Furthermore, comparative analysis against traditional methods highlights the superior performance of our models, affirming the potential of solar-powered data centers as a sustainable and environmentally responsible solution.

Designing a multi-objective energy management system in multiple interconnected water and power microgrids based on the MOPSO algorithm

In this paper, a method of the energy management system (EMS) in multiple microgrids considering the constraints of power flow based on the three-objective optimization model is presented. The studied model specifications, the variable speed pumps in the water network as well and the storage tanks are optimally planned as flexible resources to reduce operating costs and pollution. The proposed method is implemented hierarchically through two primary and secondary control layers. At the primary control level, each microgrid performs local planning for its subscribers and energy generation resources, and their excess or unsupplied power is determined. Then, by sending this information to the central energy management system (CEMS) at the secondary level, it determines the amount of energy exchange, taking into account the limitations of power flow. Energy storage systems (ESS) are also considered to maintain the balance between power generation by renewable energy sources and consumption load. Also, the demand response (DR) program has been used to smooth the load curve and reduce operating costs. Finally, in this article, the multi-objective particle swarm optimization (MOPSO) is used to solve the proposed three-objective problem with three cost functions generation, pollution, and pump operation. Additionally, sensitivity analysis was conducted with uncertainties of 25% and 50% in network generation units, exploring their impact on objective functions. The proposed model has been tested on the microgrid of a 33-bus test distribution and 15-node test water system and has been investigated for different cases. The simulation results prove the effectiveness of the integration of water and power network planning in reducing the operating cost and emission of pollution in such a way that the proposed control scheme properly controls the performance of microgrids and the network in interactions with each other and has a high level of robustness, stable behavior under different conditions and high quality of the power supply. In such a way that improvements of 41.1%, 52.2%, and 20.4% in the defined objective functions and the evaluation using DM, SM, and MID indices further confirms the algorithm's enhanced performance in optimizing the specified objective functions by 51%, 11%, and 5.22%, respectively.

Design and eco-technoeconomic analysis of a natural gas cogeneration energy management center (EMC) with short-term thermal storage

This work proposes a non-islanded cogeneration energy management center (EMC) that can be used to displace grid-level natural gas turbine systems and natural gas combustion systems for heat. The design of the proposed EMC included a weighted multi-objective optimization aimed at minimizing: i) natural gas consumption; ii) capital costs; iii) utility costs; and iv) unmet thermal demand. The decision variables consisted of the existence and capacity of the equipment comprising the EMC, including: i) a natural gas boiler; ii) an internal combustion engine that generates heat and electricity; and iii) a hot water thermal storage system. Four resulting candidates EMC designs were then compared with the status-quo (SQ) in an eco-technoeconomic analysis; The SQ draws electricity from the grid and heating for dwellings come from natural gas boilers. Emissions at grid level change which alternative is favored. The findings showed that, for a system that serves 4–5 dense urban city blocks over a 20-year lifetime, the SQ system had cumulative levelized costs of 9.6 million USD for the final consumer, while the levelized costs of the EMC designs ranged from 12.9 to 15.1 million USD. In terms of emissions, the SQ emitted 959 tonnes of CO2eq per year, while the EMC system produced around 500 tonnes of CO2eq per year depending on the year, yielding a CCA varying between 364 and 653 USD/tonneCO2eq

Optimal Data Center Energy Management With Hybrid Quantum-Classical Multi-Cuts Benders' Decomposition Method

Optimal Data Center Energy Management With Hybrid Quantum-Classical Multi-Cuts Benders' Decomposition Method

Spatio-temporal load migration potential of data centers: Evaluation and application

The wide distribution of data centers and the delay tolerance of computing tasks endow data center loads with adjustable characteristics in both temporal and spatial dimensions. Due to the characteristics of abundant spatiotemporal flexibility, data centers can participate in the optimization of power system operation and regulation. To quantify this flexibility, this paper proposes a spatiotemporal load migration potential evaluation model. Meanwhile, a data center energy management strategy is combined with the characteristics of spatiotemporal load migration, which deeply exploits the migration potential through the spatiotemporal redistribution of delay-tolerant tasks, as well as server ON/OFF scheduling and CPU operating frequency scaling across different spatial locations. A case study demonstrates that adopting the proposed approach considering an energy management strategy can effectively improve the load migration potential of data centers. The migration characteristics of data centers have great application prospects in reducing carbon emissions and enhancing operational flexibility.

Centralized Energy Management Scheme for Grid Connected DC Microgrid

Energy management in DC microgrid is complex and challenging due to the stochastic nature of renewable energy sources and load demand. Coping with the deficit power, peak demand, and power converter control operations are a few major concerns. The photovoltaic (PV) system and battery energy storage system (BESS) utilization need special attention for the reliable and efficient operation of the DC microgrid. Hence, this article proposes the centralized energy management scheme (CEMS) in the DC microgrid to address the abovementioned challenges. The CEMS coordinates among the various physical layer components and computational layer (i.e., control layer) of the DC microgrid. It utilizes the multioptimization for energy management during the peak and <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">off</small> -peak load demand, optimal usage of PV-BESS, and performs the optimal load shedding operation. Furthermore, an optimal utilization of the PV-BESS problem is formulated and solved using linear programming to avoid excess usage of AC grid power. In deficit PV power, the optimal load shedding operation is formulated as mixed-integer linear programming problem to ensure the DC bus voltage regulation and power balance. The proposed CEMS demonstrates the effective operation of the PV system in maximum power point tracking, <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">off</small> -MPPT mode (i.e., voltage control), and BESS's charging/discharging operation. It is showcased that the hierarchical control structure of CEMS improves the performance, efficiency, and reliability of the DC microgrid operation. Hence, a 48 V and 1.2 kW PV-BESS-based DC microgrid system is developed to show the efficacy of the proposed algorithm under different scenarios in MATLAB/Simulink. The real-time simulator results of the developed system, obtained using OPAL-RT OP4510, confirmed the effectiveness and reliability of the proposed scheme.

Multi-agent based energy community cost optimization considering high electric vehicles penetration

This paper proposes a multi-agent based solution to minimize the energy cost of an energy community, in day-ahead, that includes high penetration of electric vehicles. The proposed approach employs a structure of agents, including a central coordinator and energy management agents. To minimize the energy cost and to optimize the energy balance of the energy community, the proposed approach considers the energy demand and supply (i.e., photovoltaic generation), the battery storage systems’ charge and discharge actions, and the charging and discharging schedules of electric vehicles, including the possibility of vehicles to charge and discharge energy in public charging stations. The optimization problem is formulated as a mixed-integer linear programming problem, which is solved by an open-source solver and compared with the use of a commercial solver. The simulation results show that the proposed day-ahead approach can significantly reduce the cost of the energy community while ensuring a reliable and stable operation. Comparing the proposed solution with a centralized implementation, it is possible to significantly reduce the optimization time from hours to few seconds. Overall, the proposed multi-agent based solution provides a promising solution for the optimization of energy communities with high penetration of electric vehicles.

Joint optimization of energy trading and consensus mechanism in blockchain-empowered smart grids: a reinforcement learning approach

Under the trend of green development, the traditional fossil fuel and centralized energy management models are no longer applicable, and distributed energy systems that can efficiently utilize clean energy have become the key to research in the energy field nowadays. However, there are still many problems in distributed energy trading systems, such as user privacy protection and mutual trust in trading, how to ensure the high quality and reliability of energy services, and how to motivate energy suppliers to participate in trading. To solve these problems, this paper proposes a blockchain-based smart grid system that enables efficient energy trading and consensus optimization, enabling electricity consumers to obtain high-quality, reliable energy services and electricity suppliers to receive rich rewards, and motivating all parties to actively participate in trading to maintain the balance of the system. We propose a reputation value assessment algorithm to evaluate the reputation of electricity suppliers to ensure that electricity consumers receive quality energy services. To minimize the cost, maximize the benefit for the electricity suppliers and optimize the system, we present an algorithm based on reinforcement learning DDPG to determine the power supplier, power generation capacity, and consensus mechanism between nodes to obtain power trading rights in each round. Simulation results show that the proposed energy trading scheme has good performance in terms of rewards.

Adaptive centralized energy management algorithm for islanded bipolar DC microgrid

Abstract In this paper, an adaptive centralized energy management algorithm is developed for an islanded bipolar DC microgrid (BDCMG). The proposed microgrid consists of a solar photovoltaic (SPV) system, a wind energy conversion system (WECS), and a battery energy storage system (BESS). The SPV and WECS are operated in maximum power point tracking mode to extract maximum power. The adaptive control algorithm focus on power sharing at a bipolar DC bus based on variations in wind speed, solar irradiance, battery state of charge (SoC), and load. Furthermore, this algorithm maintains voltage balance and voltage regulation on a bipolar DC bus without the use of a communication link and provides effective energy management through coordinated control of renewable energy sources and BESS. The efficacy of the proposed bipolar DC microgrid is simulated and validated on a lab scale prototype to verify its feasibility, and it is observed that the results are found to be satisfactory.

Advanced data analytics modeling for evidence-based data center energy management

Over the past few decades, the demand for Data Center (DC) services has significantly increased due to the world’s growing need for internet access, social networking, and data storage. Data Centers are among the most energy-intensive businesses, so optimizing IT operations in DC requires energy-efficient techniques. This paper presents AI based modeling strategies for effective energy management with a particular emphasis on DC’s two most energy intensive systems (i.e., cooling and IT systems). This study addresses the issues of IT equipment performance degradation, inappropriate IT room thermal conditions, inefficient workload placement, and excessive energy waste. This research entails the application of machine learning for DC thermal classification, and deployment of deep learning models to predict resource utilization and energy consumption in DC. Furthermore, a comparative analysis is performed with existing relevant methods to demonstrate the effectiveness and accuracy of the proposed AI techniques. The findings of this study also provide evidence-based recommendations for DC efficient energy management.

A new rule-based self-reconfigurable energy management control of grid and batteries connected hybrid regenerative solar power system for critical loads

Renewable energy sources are increasing and diversifying day by day. For this reason, new and more efficient controller designs are being developed on diversified renewable energy sources. Thus, renewable energy power systems can vary with the development of both topological, power system element controllers and energy management system controllers in line with the desired demands. In this paper, Hybrid Regenerative Power System (HRPS) can operate at low power with a designing a novel Central Energy Management System (CEMS) that is suitable for the critical AC & DC loads and it is efficient thanks to the regenerative effect. Equipments in this power system were firstly created and they are connected each other to obtain HRPS at MATLAB-Simulink environment. After the HRPS topology emerged, power flow scenarios are automatically operated according to the proposed Energy Management System (EMS) on the HRPS topology with using real irradiance and temperature spectrums. Then, according to the consequences obtained, how efficient the power system is in itself was calculated and when compared to the literature, it was emphasized in which cases it was advantageous compared to other EMS studies.

Reward Shaping-Based Actor–Critic Deep Reinforcement Learning for Residential Energy Management

Residential energy consumption continues to climb steadily, requiring intelligent energy management strategies to reduce power system pressures and residential electricity bills. However, it is challenging to design such strategies due to the random nature of electricity pricing, appliance demand, and user behavior. This article presents a novel reward shaping (RS)-based actor–critic deep reinforcement learning (ACDRL) algorithm to manage the residential energy consumption profile with limited information about the uncertain factors. Specifically, the interaction between the energy management center and various residential loads is modeled as a Markov decision process that provides a fundamental mathematical framework to represent the decision-making in situations where outcomes are partially random and partially influenced by the decision-maker control signals, in which the key elements containing the agent, environment, state, action, and reward are carefully designed, and the electricity price is considered as a stochastic variable. An RS-ACDRL algorithm is then developed, incorporating both the actor and critic network and an RS mechanism, to learn the optimal energy consumption schedules. Several case studies involving real-world data are conducted to evaluate the performance of the proposed algorithm. Numerical results demonstrate that the proposed algorithm outperforms state-of-the-art RL methods in terms of learning speed, solution optimality, and cost reduction.

Energy management system in smart buildings based coalition game theory with fog platform and smart meter infrastructure

This paper proposes a central energy management system (EMS) in smart buildings. It is based on the coalition method for optimal energy sharing between smart buildings. Game theory is applied to obtain an optimal allocation of the building's surplus energy on the deficient energy buildings using the Shapley value, which enables the unequal energy distribution based on the energy demand. The main objective is reducing energy waste while preserving the generation/demand balance. The fog platform with memory storage is applied, which handles all the measured data from the smart buildings through Wi-Fi-based communication protocol and performs the EMS program. The smart meter links the smart buildings with the fog-based EMS central unit. Two scenarios are implemented based on the difference between total deficient and surplus energy. Coalition game theory is applied for optimal surplus energy allocation on deficient buildings when the total energy surplus is lower than the total energy deficient. Also, there is a one-to-one relationship between the surplus and deficient building; if the surplus energy is larger than the deficit, the extra surplus energy is stored for further usage. The proposed EMS is applied and tested using a smart city with 10 buildings in the MATLAB program. A comparison between the result obtained with and without applying the proposed method is performed. The performance of the fog platform is introduced based on the run and delay time and the memory size usage. The results show the effectiveness of the proposed EMS in a smart building.

RECLAIM: Renewable Energy Based Demand-Side Management Using Machine Learning Models

The diesel generators sets (DGs) and battery storage systems (BSS) are the essential energy sources in a modern high-rise buildings. In this paper DG, BSS and Photovoltaic system (PV) has been considered to minimize the grid power injection using a centralized Energy Management System (EMS). Machine Learning (ML) techniques are used to predict the performance of various regression models by comparing grid power and load curves. It includes Artificial Neural Network (ANN), Wide Neural Network (WNN), Linear Regression (LR), Linear Regression Interaction (LR-I), Linear Regression Stepwise (LR-S), Regression Fine Tree (RF-T), Regression Coarse Tree (RC-T) and Gaussian Process Regression (GPR) based techniques. The Demand Side Management (DSM) techniques such as peak shaving and valley filling is integrated with ML technique in a Hybrid energy source (HS) system.The comparative analysis of results depicts the effective reshaping of the grid profile without scheduling or disconnecting the loads. Matlab simulation software is used to validate the results.

Energy management strategy for standalone DC microgrid system with photovoltaic/fuel cell/battery storage

Standalone DC microgrids often have challenges in energy management for a long time horizon due to uncertain renewable energy sources and volatile loads. This paper presents a centralized energy management strategy(EMS) for a standalone DC microgrid with solar PV, fuel cells, and a battery energy storage system (BESS). The proposed EMS method is designed to improve the longevity of BESS, reliability, and reduce the hydrogen intake. In the proposed EMS, the PV system de-rating method is used to overcome the deep charging of battery under low-demand conditions. The fuel cell power supply is varied using a reverse sigmoidal function of the Batterys state of charge (SoC). This improves the hydrogen fuel efficiency and also helps in minimizing deep discharge of the battery under heavy loading conditions. The centralized EMS is fed with load power, battery SoC, and individual source power information. Consequently, the EMS provides decisive commands to the individual source local controller to control the respective output power. The efficacy of the proposed EMS under multiple operating conditions is evaluated in both simulation environment and on a hardware prototype of a DC microgrid.

A Stackelberg Game Theory Based Demand Response Algorithm for Domestic Consumers

A stackelberg game algorithm is proposed to optimize real time demand response equilibrium between consumers and utilities with the integration of renewable energy sources. Stackelberg game is played between one leader and N followers to provide interaction between utility and consumers in the presence of energy management center. The objective of consumers is to minimize their energy tariffs by scheduling their appliances. Appliances are categorized as non-shiftable, shiftable and curtailable appliances based on their applications. The objective of the utility is to maximize its profit by managing peak to average energy consumption ratio, penetration of renewable energy sources as a primary source and during peak hours, integrate electric vehicles as a backup. The existence of a unique stackelberg equilibrium that provides an optimal energy consumption scheme to each appliance at the consumer end was demonstrated. The simulation results showed the optimal real time demand response using stackelberg game algorithm is effective for achieving the optimal load control.

Tree-Based Ensemble Machine Learning Techniques for Power System Static Security Assessment

Emerging machine learning techniques (MLTs) have proven to be effective in many applications including power system studies. In this paper, the performance of tree-based ensemble MLTs has been investigated to assess the static security of power system and compared with base-line methods. System wide composite security index is used to assess the severity of a contingency against bus voltage and power flow violations. The scope of study is limited to widely used base line estimators: multilayer perceptron model, support vector machine, K-nearest neighbors, decision tree and ensemble-based estimators: random forest, AdaBoost, gradient boost, and extra tree. Performance analysis is done in terms of accuracy, classification rate and training time. To further investigate the suitability of these approaches to time-constrained security environment at Energy Management Center, recursive feature elimination (RFE) method is used to contain the curse of dimensionality and performance evaluation is done on test systems extending from IEEE 14-bus to IEEE 118-bus including practical 75-bus Indian system. Ensemble-based methods in general and ETC in particular outperformed base line methods in terms of accuracy and computational time with RFE.

Optimal stochastic scheduling of plug-in electric vehicles as mobile energy storage systems for resilience enhancement of multi-agent multi-energy networked microgrids

This paper presents an optimal scheduling of plug-in electric vehicles (PEVs) as mobile power sources for enhancing the resilience of multi-agent systems (MAS) with networked multi-energy microgrids (MEMGs). In each MEMG, suppliers, storage, and consumers of energy carriers of power, heat, and hydrogen are taken into account under the uncertainties of intermittent nature of renewable units, power/heat demands, and parking time of PEVs. In the case of contingencies, the proposed algorithm supplies energy to the on-fault MEMGs from normal-operated grid-connected MEMGs, using mobile PEVs. The procedure of selecting PEVs to supply energy to the on-fault MEMGs is performed in three stages. Initially, both on-fault and normal-operated MEMGs inform the central energy management system (EMS) about the amount of required energy and the amount of available energy from existing PEVs. Further, central EMS prioritizes the MEMGs among networked MEMGs to supply the energy support to the on-fault islanded MEMG. Lastly, the chosen MEMGs select their available efficient PEVs to supply energy to the on-fault islanded MEMG. Considering two diverse faulty case studies, the proposed technique is investigated in a MAS with four networked MEMGs. Simulated results demonstrate that the proposed algorithm enhances the resilience of MEMGs (over 25%) even without a physical connection between the MEMGs.

Smart centralized energy management system for autonomous microgrid using FPGA

Smart energy management system (SEMS) has become an effective energy-saving tool. In this paper, an efficient energy management system is used for a hybrid system consists of PV, Fuel Cells (FCs), and wind energy systems. The hybrid system is accompanied by a battery energy storage system to act as a backup source in case that the loads exceed the power produced from the three sources. The SEMS is implemented on Altera Cyclone IV EP4CE6 field-programmable gate arrays (FPGA) board using QUARTUS II software and LABVIEW. FPGAs are highly flexible and reprogrammable logic devices. Analog to digital converter (ADC) is used to interface the current and voltage of the three sources with FPGA. As a result of the weather fluctuations, the voltage of the three sources is not stable. Therefore, the voltage of the three sources is adjusted using the Buck-Boost converter technology, which is set at 24 V until the merging process takes place. The results prove the high efficiency of the FPGA and its high response to the load changes.

Performance of a Mid-Size Net-Zero Energy Solar House

The University of Alabama at Birmingham (UAB) was one of 16 collegiate teams from around the world that participated in the U.S. Department of Energy Solar Decathlon 2017 competition. An interdisciplinary team of students from across the university was engaged in a 2-year long process to design and build a house that is powered completely by solar power. The house was equipped to run all the typical appliances of an average modern house at similar levels on a conventional power grid. The net-zero house was built and tested on the UAB campus. Considering Birmingham’s weather, a safe room was built to ensure the safety of occupants during events of extreme weather, such as a tornado. A ductless HVAC unit consisting of an inverter-type 3-speed outdoor unit supplied refrigerant to four high-wall indoor units providing the primary source of space conditioning. To achieve a model of efficiency and cost effectiveness, the house incorporated a heavily insulated envelope and precise glazing protection. The roof, floor framing and walls had R-30 batt and foam insulation. With a design informed by southern vernacular language, the building is oriented to maximize solar access and to use roof planes for shading the majority of the year. Peak power generation of the panels was recorded at 9.6 kW. The home has a centralized energy management system that can provide access to energy consumption data and allow control of lighting, appliances, and plug loads remotely. Energy modeling showed that the annual electricity consumption for heating and cooling with variation in wall types were 8470 to 11,661 kWh. For the month of October, it was calculated varying from 683 to 763 kWh, with varying air changes per hour from 0 to 1.5.