Multi-Agent Reinforcement Learning based Distributed Renewable Energy Matching for Datacenters

SUMMARYIn the future, renewable energy (RE) generation systems will be constructed and interconnected to the power system by nonutility entities such as generation companies and residential customers. Therefore, generation expansion planning (GEP) should consider the effects of RE generation. The purpose of this paper is to discuss the best GEP from the economic, supply reliability, and environmental aspects considering the penetration of RE generation. For this purpose, this paper proposes a model of a utility's GEP process considering RE generation. The proposed method is based on stochastic dynamic programming (SDP) in which long‐term uncertainties are modeled by geometric Brownian motion (GBM) and the binomial lattice process. The variation of RE generation output due to the weather conditions (short‐term uncertainty) is also considered in the proposed method by means of the net load curve.

China is faced with significant challenges in simultaneously promoting rural development and reducing carbon emissions. However, the issue of quantifying and addressing carbon emissions in rural areas has not been adequately addressed. Accurately quantifying these emissions is crucial for developing effective strategies to reduce carbon output. In this study, the historical evolution and spatial distribution of rural carbon emissions in northwestern China from 2006 to 2019 were evaluated across five key sectors: residential energy consumption, agricultural machinery, solid waste management, planting practices, and breeding industry activities. During this period, total carbon emissions in rural areas of northwest China steadily increased from 60.15Mt to 83.49Mt at an annual growth rate of 2.55%. Given the complex interplay between economic and social factors driving these changes, the future trajectory of rural carbon emissions remains uncertain. To analyze the underlying drivers behind regional variations in carbon emissions over time, we constructed an LMDI model which revealed that economic growth primarily contributed to regional increases in carbon output. Furthermore, due to a remarkable annual growth rate of 35.17% in renewable energy generation (such as photovoltaic and wind power), it can be inferred that if renewable electricity were included within our calculations for carbon emission statistics, northwest China’s rural areas achieved a state of being effectively ‘carbon-neutral’ by 2019 solely from a production-based perspective.

Berkeley Lab’s annual status report on U.S. renewables portfolio standards (RPS) provides an overview of key trends associated with U.S. state RPS policies. The report, published in slide-deck form, describes recent legislative revisions, key policy design features, compliance with interim targets, past and projected impacts on renewables development, and compliance costs. The 2017 edition of the report presents historical data through year-end 2016 and projections through 2030. Key trends from this edition of the report include the following: -Evolution of state RPS programs: Significant RPS-related policy revisions since the start of 2016 include increased RPS targets in DC, MD, MI, NY, RI, and OR; requirements for new wind and solar projects and other major reforms to the RPS procurement process in IL; and a new offshore wind carve-out and solar procurement program in MA. -Historical impacts on renewables development: Roughly half of all growth in U.S. renewable electricity (RE) generation and capacity since 2000 is associated with state RPS requirements. Nationally, the role of RPS policies has diminished over time, representing 44% of all U.S. RE capacity additions in 2016. However, within particular regions, RPS policies continue to play a central role in supporting RE growth, constituting 70-90% of 2016 RE capacity additions in the West, Mid-Atlantic, and Northeast. -Future RPS demand and incremental needs: Meeting RPS demand growth will require roughly a 50% increase in U.S. RE generation by 2030, equating to 55 GW of new RE capacity. To meet future RPS demand, total U.S. RE generation will need to reach 13% of electricity sales by 2030 (compared to 10% today), though other drivers will also continue to influence RE growth. -RPS target achievement to-date: States have generally met their interim RPS targets in recent years, with only a few exceptions reflecting unique state-specific policy designs. -REC pricing trends: Prices for renewable energy certificates (RECs) used to meet general RPS obligations fell in most markets in 2016, as surplus RPS supplies emerged in many regions. Price trends for solar RECs were more varied, with a particularly pronounced drop in MD. -RPS compliance costs and cost caps: RPS compliance costs totaled $3.0 billion in 2015 (the most-recent year for which relatively complete data are available), which equates to 1.6% of average retail electricity bills in RPS states. Though total U.S. RPS compliance costs rose from 2014, future cost growth in most RPS states will be capped by cost containment mechanisms.

In the context of increasing sustainability through renewable energy utilization in Saudi Arabia, this study was motivated by the need to understand the impact of COVID-19 on gym-goers’ attitudes and behaviors towards renewable energy generating using their human kinetic energy (HKE). A comparative analytical study was conducted using a pre-COVID-19 survey (n = 96) and a post-COVID-19 survey (n = 385) of gym-goers in Saudi Arabia. The surveys consisted of closed-ended Likert-type questions that measured participants’ attitudes, behaviors, and willingness related to renewable energy and HKE generation. The results of the Chi-square test of independence showed that there was no significant difference in the participants’ attitudes towards behaviors and awareness regarding renewable energy and HKE generation between the pre-COVID-19 and post-COVID-19 surveys. However, the study revealed an overall promising level of support for sustainable energy and HKE generation among gym-goers. This led to the conclusion that there is potential for the installation of sustainable HKE generation systems in gyms. The main implication of this finding is the need for greater education and awareness raising campaigns as well as incentives to support the implementation of renewable energy solutions. The research contributes to sustainable development and Saudi Vision 2030 by exploring potential opportunities for increasing the share of renewable energy in the overall energy mix and promoting sustainable development.

Towards renewable energy generation and low greenhouse gas emission in high-income countries: Performance of financial development and governance

High energy consumption in data center has become a key problem, and the Carbon (CO $$_2$$ ) emissions from data center have serious impacts on environment. To improve traditional performance-energy model and minimize energy consumption in cloud data center, an energy-efficient and QoS-aware (EEQoS) model is developed for virtual resources consolidation. The EEQoS model is designed to investigate the trade-off between energy and QoS. The user’s satisfaction degree and Euclidean distance of resources are also respectively applied to the QoS model and power model. In addition, this EEQoS model is combined with the particle swarm optimization algorithm by setting the power consumption per QoS value as the objective function to consolidate virtual resources in data center. Compared with traditional power model, the proposed model reduces response time by 27.2%, cost by 31.4%, SLA violations by 40.5% and improves throughput by 13.5% on average while increasing less than 3.8% energy consumption during the whole test period.

Unveiling the dynamic impact of energy generation on economic sustainability in Canada: A roadmap towards sustainable development

This article examines the dynamic relationships among output, carbon emission and renewable energy generation of India and China during the period 1972 to 2011 using a multivariate vector error correction model (VECM). The results for India reveal unidirectional short-run causality from carbon emission to renewable energy generation and from renewable energy generation to output, whereas in the long run, the variables have bidirectional causality. Causalities in China give a rather different scenario, with a short-run unidirectional causality from output to renewable energy and from carbon emission to renewable energy generation. In the long run, for China, unidirectional causality is found from output to renewable energy generation, while bidirectional causality is found between carbon emission and renewable energy generation.

Between the end of 2015 and mid-2016, Argentina enacted a new renewable energy (RE) promotion regime and launched the so-called RenovAr Program, a public plan that aimed to increase the development of RE generation projects through a series of international bids for power capacity. The proposed goal was to move from 2% RE generation of the total Argentina’s energy mix in 2015, to 20% by 2025. The RenovAr Program, along with RE promotion policies, succeeded in tripling RE generation capacity by 2019. However, over the last years the pace of new RE capacity incorporation has significantly slowed down, making it unlikely to reach the 20% target by 2025. This work analyses the positive aspects of the Argentinean RE promotion regime and the RenovAr Program that led to its rapid success, and it considers the main weaknesses that have prevented the further expansion of RE capacity. This paper concludes that, while designing policies to attract international funding for RE projects is beneficial for increasing RE generation capacity, ensuring its continuous growth in the long-term also requires addressing macroeconomic challenges and sustainable political support.

Synthetic inertia and frequency support assessment from renewable plants in low carbon grids

In the presence of latency sensitive geo-distributed applications, users require fast service for their queries. Cloud computing provides physical servers from its data center in order to process user requests. The cloud data center consumes a huge amount of energy due to lack of management of the data center servers as the container-based service consolidation is a nontrivial task. Since the containers require less resource footprint, consolidating it in servers might make resource availability sparse. In order to reduce the energy consumption of the cloud data center, we have proposed a green container-based consolidation of the services so that the maximum number of servers can be put into idle mode without affecting the application quality of experience. The service consolidation problem has been formulated as an optimization problem considering minimization of total energy consumption of the data center as the objective, and an algorithm named Energy Aware Service consolidation using baYesian optimization (EASY) has been proposed to solve the optimization. We have evaluated the EASY algorithm in simulation using python. The experimental results have shown that EASY improves the total energy consumption of the data centers. This improvement comes at the cost of a small increase of service response time as there exists a trade-off between energy consumption and service response time.

Cloud data centers traditionally use grid energy which emits carbon dioxide and dissipates heat to the environment. To decrease the amount of carbon emissions, many Cloud Service Providers (CSPs) are exploring the possibility of using renewable energy to operate their data centers showing a clear trend to migrate towards green cloud datacenters. Green cloud data centers typically use solar and wind as renewable energy sources. To meet the highly dynamic user demand and to ensure effective power management, an efficient methodology is proposed that performs various load scheduling and power management plans with the integration of renewables for geographically distributed green cloud datacenters. The proposed approach applies nature-inspired Modified Bacterial Foraging Optimization Algorithm (MBFOA) to balance load and reduce eco-aware energy cost. Simulations were conducted to determine the efficacy of the proposed MBFOA algorithm. The algorithm efficiently reduces eco-aware power cost of the system than the baseline methods.

As renewable energy (RE) rapidly integrates into society to meet the growing demand for affordable and clean energy (SDG 7, United Nations’ Sustainable Development Goal 7), it is crucial to analyze the merits and flaws of renewable energy generation by assessing its impact on the global economy and social wellbeing. This paper performs an investigative study on the correlation and causality between renewable energy, economy, and environmental indicators. The data was gathered from diverse sources, including The Center for Climate and Energy Solutions, National Aeronautics and Space Administration Goddard Institute for Space Studies (NASA GISS), and Our World in Data, for the temperature, carbon dioxide (CO2) emissions, and gross domestic product (GDP) data. Significant Granger p-values were obtained for RE generation’s ability to forecast CO2 emissions and temperature, while discovering a strong positive correlation between CO2 and RE generation. The findings revealed that RE has limited effects on the global economy but has considerable implications on social and ecological well-being.

This paper proposes a privacy-aware collaborative energy management framework for cloud data centers (CDCs) operating within an electricity market environment. Each CDC's energy agent (CDCEA) minimizes the total CDC's energy cost by optimizing its energy profile in response to dynamic locational marginal prices and the nodal carbon intensity of the power system issued by the independent system operator (ISO), who clears the energy market. CDCEAs integrate the on-site energy resources of CDCs, such as renewable energy and energy storage systems, and leverage the spatiotemporal flexibility of computing workloads through coordination with cloud service providers (CSPs). CSPs manage privacy-preserving workload scheduling and resource allocation by collaborating with associated CDCEAs and with other CSPs, while ensuring quality of service and adherence to service-level agreements. A two-stage optimization model is developed to coordinate decision-making between the ISO and CDCEAs in energy market in the first stage, and to address the coordination of CSPs with associated CDCEAs and other CSPs in the second stage. The proposed two-stage model is solved using a three-nested distributed iterative algorithm that combines privacy-aware hierarchical and peer-to-peer ADMM methods with an analytical target cascading (ATC) approach. The method is validated using a real-world case study in Northern Virginia, USA, demonstrating its effectiveness in reducing costs and enhancing sustainability.

In this paper, a cost minimization problem is formulated to intelligently schedule energy generations for microgrids equipped with unstable renewable sources and energy storages.
\nIn such systems, the uncertain renewable energy will impose unprecedented scheduling challenges. To cope with the fluctuate nature of the renewable energy, an uncertainty model based on renewable energies’ moment statistics is developed. Specifically, we obtain the mean vector and second-order moment matrix according to predictions and field measurements and then define
\nuncertainty set to confine the renewable energy generation. The uncertainty model allows the renewable energy generation distributions to fluctuate within the uncertainty set. We develop chance constraint approximations and robust optimization approaches based on a Chebyshev inequality framework to firstly transform and then solve the scheduling problem. Numerical results based on real-world data traces evaluate the performance bounds of the proposed scheduling scheme. It is shown that the temporal correlation information of the renewable energy within a proper time span can effectively reduce the conservativeness of the solution. Moreover, detailed studies on the impacts of different
\nfactors on the proposed scheme provide some interesting insights which shall be useful for the policy making for the future microgrids.