Data Center Energy Research Articles

Optimizing data center energy consumption via energy complementarity scheduling

Considering the degree of flexible energy supply and demand matching, four types of energy sources were combined and a dynamic multi-energy combination optimization scheduling model was established for data centers to realize the optimal scheduling of multiple energy sources to meet the minimum overall energy consumption of data centers. Model construction and simulation analysis were used to conduct the research. To maximize the use of wind power and photovoltaic renewable energy, a multi-objective optimization scheduling model was constructed with the objectives of maximum flexibility of supply and demand matching in a single day, minimum penalty for predicted photovoltaic power consumption, minimum average service time of data centers, and minimum overall energy consumption of data centers. The NSGA-III framework was used to solve the constructed model. The results showed that even in areas rich in renewable energy, because of the limited power of renewable energy generation, the purchase of electric energy and natural gas as the main energy source remained necessary to ensure the stable operation of data centers. Data centers can use the complementarity of photovoltaic and wind power generation times to achieve basic complementarity of power generation. In terms of cooling energy consumption, natural gas waste heat-driven absorption refrigerator output accounted for a significant proportion, approximately twice that of the electric refrigerator and could effectively reduce the energy consumption of electric refrigeration. The simulation results of energy scheduling model show that it is necessary to increase the proportion of natural gas and green energy in data center energy supply, maximize the use of renewable energy such as wind power and photovoltaic power generation, give full play to the characteristics of wind power and photovoltaic power generation time complementarity, and build a green energy complementary data center energy supply mode.

Techno-economic analysis of combined photovoltaic cells and hydrogen energy systems for data center energy consumption

Techno-economic analysis of combined photovoltaic cells and hydrogen energy systems for data center energy consumption

A novel multi-modal Federated Learning based thermal-aware job scheduling framework

Cooling costs constitute more than half of the total data center energy expenditure. Thermal imbalance results in hotspot regions requiring additional cooling power. To reduce it, thermal aware job scheduling is a well-known software solution that is subject to predicting correct server temperatures. Existing solutions have not explored intelligent solutions and rely only on logic based algorithms to allocate tasks that work on predefined rules. Few deep learning based solutions that are proposed, have not explored its alternatives and existing data modalities in data centers, resulting in inefficient models. Existing literature only proposes solutions based on unimodal tabular data. Therefore, we propose a multimodal architecture that considers different underlying data modalities in data centers to increase the model’s efficiency and predict correct server temperatures. The increasing production of data and the need for storage and processing units has led to the development of distributed data centers. Existing techniques are limited to individual data centers which fail to consider the data privacy restrictions that arise while dealing with distributed scenarios. Findings from our simulations affirm our proposed scheme in terms of the objectives mentioned above. We propose a federated learning architecture that efficiently deals with distributed data centers while ensuring privacy. Our simulation results show an overall increase in the efficiency of the model in comparison to an existing intelligent solution. Furthermore, we provide comparative results that show how our model performs better and achieves lower thermal imbalance as compared to an existing scheme.

Experimental investigation on the application of cold-mist direct evaporative cooling in data centers

The rapid development of the internet era has driven the construction of numerous data centers worldwide. In hot climate, data centers consume significant energy for cooling. Cold-mist direct evaporative cooling offers a natural cooling solution that can help reduce this energy consumption. This study investigates the temperature and relative humidity changes of natural high-temperature air after being cooled using a cold-mist direct evaporative cooling test bench. The effects of several control factors such as spray angle, spray flow rate, and air speed on the cold-mist direct evaporative cooling performance was systematically examined. The findings revealed that: (1) the optimal spray angle for cold-mist direct evaporative cooling treatment air is 65°; (2) high-temperature air between 27 °C and 37 °C can be cooled to 23.57 °C – 25.58 °C after cold-mist direct evaporative cooling treatment, with relative humidity levels of 67.0 % – 78.8 %, meeting the air supply requirements for data centers; (3) the proposed approach could reduce the data center energy consumption by 14 % – 41 %, while extending the annual natural cooling period by 3.16 % – 20.45 %.

Next-generation data center energy management: a data-driven decision-making framework

In the era of society’s ongoing digitization and the exponential growth in data volume, alongside a growing energy demand, energy management plays an integral role in data centers (DCs) and is a key factor in the quest for decarbonization. In light of the complex nature of DCs, traditional energy management strategies are inadequate. This research introduces a data-driven decision-making framework for DCs, grounded in the OODA (Observation, Orientation, Decision, and Action) loop and based on insights from an Ericsson-operated DC in Linköping, Sweden. The developed framework enables DCs to enhance energy efficiency effectively. Rooted in the OODA loop and leveraging extensive datasets from DCs’ building management systems, this framework aids in decreasing cooling energy usage through strategic, data-driven decision-making. By adopting AI methods, specifically K-means clustering in this research, for continuous monitoring and fine-tuning (Proportional, Integral, Derivative) PID parameters, the framework aids in improving operational efficiency.

Real-Time Energy Management Method for Data Center Considering Shiftable Workload and Renewable Energy

Aim: This study aimed to improve the electricity economy of data centers. Background: Data centers have been rapidly developed over the last decade to balance increasing computing workloads. Such a rapid development has been found to result in a significant increase in higher electricity expenses for data centers. Objective: The objective of this study was to improve the operation efficiency of a data center by developing a data center energy management model considering shiftable data workloads, energy storage, and renewable energy. Method: First, a data center workload model has been established to describe its temporal shifting characteristics. Then, a data center power consumption model has been developed to describe the energy usage of IT equipment and cooling systems. Finally, the energy management model for the data center has been developed to reduce the electrical energy costs of the data center. Results: Case studies have been performed on a trial system to illustrate the validity of the proposed model, which has been found to meet the requirements. Conclusion: Rationally re-scheduling the delay-tolerant workloads and making full use of energy storage can enhance the flexibility of the power system and reduce electricity costs by 19.06%.

Towards Energy-Efficient Data Centers: A Comprehensive Review of Passive and Active Cooling Strategies

With the rapid growth of cloud computing, the number of data centers (DCs) continuously increases, leading to a high-energy consumption dilemma. Cooling, apart from IT equipment, represents the largest energy consumption in DCs. Passive design (PD) and active design (AD) are two important approaches in architectural design to reduce energy consumption. However, for DC cooling, few studies have summarized AD, and there are almost no studies on PD. Based on existing international research (2005-2024), this paper summarizes the current state of cooling strategies for DCs. PD encompasses floors, ceilings, and layout and zoning of racks. Additionally, other passive strategies not yet studied in DCs are critically examined. AD includes air, liquid, free, and two-phase cooling. This paper systematically compares the performance of different AD technologies on various KPIs, including energy, economic, and environmental indicators. This paper also explores the application of different cooling design strategies through best-practice examples and presents advanced algorithms for energy management in operational DCs. This study reveals that free cooling is widely employed, with Artificial Neural Networks emerging as the most popular algorithm for managing cooling energy. Finally, this paper suggests four future directions for reducing cooling energy in DCs, with a focus on the development of passive strategies. This paper provides an overview and guide to DC energy-consumption issues, emphasizes the importance of implementing passive and active design strategies to reduce DC cooling energy consumption, and provides directions and references for future energy-efficient DC designs.

Data Center Power and Energy Management: Past, Present, and Future

Data Center Power and Energy Management: Past, Present, and Future

Energy performance analysis of multi-chiller cooling systems for data centers concerning progressive loading throughout the lifecycle under typical climates

The increasing demand for cooling energy in data centers has become a global concern. Existing studies lack a comprehensive analysis of the energy performance of widely used multi-chiller cooling systems in air-cooled data centers throughout their lifecycle, especially concerning progressive loading. To bridge this gap, this study conducts a thorough assessment of the energy performance of multi-chiller cooling systems throughout the entire lifecycle. Additionally, the impact of climate conditions on the energy efficiency of the cooling systems is analyzed, considering design variations for typical climates. Multi-chiller cooling system models are developed using the test data of cooling equipment and typical control algorithms. The energy performance of the cooling system is thoroughly analyzed under full-range cooling loads and climate conditions. Results show that free cooling time could differ up to 1442 hours at different part load ratios in the same location. Furthermore, the cooling system’s coefficient of performance (COP) varies significantly, by up to 6, at different part load ratios, corresponding to a difference in power usage effectiveness (PUE) up to 0.14. Notably, the average cooling system COP throughout the lifecycle loading is found to be only 11.7, 2.9 lower than the design system COP.

Analysis of refrigerant/lubricant distribution and operating characteristics of a rotary booster for data center free cooling

A booster-driven loop free cooling system plays a crucial role in enhancing data center energy efficiency and reducing emissions due to the special booster with low pressure ratio as a vital driving component. And the impact of lubricant and its fraction on the booster operation are the key points under low pressure ratio, as well as lubricant distribution. In this paper, the refrigerant/lubricant two-phase mixing flow in the working chamber of a booster was investigated under low pressure ratio for data center free cooling. The volume of fluid (VOF) multiphase flow model combining the working chamber of the rotary booster and discharge valve was built up to analyze the dynamic characteristics of the booster under free cooling conditions. The effect of lubricant fraction and its working frequency on the operating indexes (exhaust temperature/pressure, mass flow rate, volumetric efficiency, indicated specific work, energy efficiency ratio) were studied. The lubricant distribution and its changes for different angles and frequencies were presented. The results indicate that it can improve the thermal performance of the booster effectively with appropriate volume fraction of lubricant. The lubricant in the two-phase flow reduces the inner temperature and pressure in the booster working chamber, which changes the exhaust pressure and volumetric efficiency as well as the cooling capacity. The exhaust pressure and the mass flow rate increase notably for the refrigerant/oil mixture with the increasing lubricant volume fraction, while other performance indexes decrease. The exhaust pressure and the mass flow rate increase by 0.51 % and 47.14 %, separately, when the lubricant volume fraction raises from 0 to 5 % at operating frequency of 50 Hz. Whereas, the exhaust temperature, refrigerant mass flow rate, volumetric efficiency, indicated specific power, cooling capacity and EER decrease by 18.11 %, 4.81 %, 47.14 %, 38.94 %, 4.81 % and 26.36 %, respectively. The investigations can support the lubrication and sealing of the booster with low pressure ratio for free cooling to reach efficient and reliable operation.

Quantum Enhanced Josephson Junction Field-Effect Transistors for Logic Applications

The need of data centers for cloud and network computing has dramatically increased power consumption. In 2014, data centers in the U.S. consumed an estimated 70 billion kWh, representing about 1.8% of total U.S. electricity consumption [“United States Data Center Energy Usage Report," published in June 2016 and supported by the Federal Energy Management Program of the U.S. Department of Energy]. Worldwide, the International Energy Agency estimates that currently about 1% of all global electricity is used by data centers. It is expected that by 2030 the electricity use of information and communications technology may exceed 20% of all global electricity. The microelectronics industry has been focusing on aggressively shrinking the size of logic and memory devices to reduce power consumption. However, this scaling is going to eventually stop as we approach the fundamental materials limit. Moreover, the power dissipation due to the increased number of transistors is also expected to be a limiting factor for processor performance.Superconducting computing has been suggested as a promising approach for low-energy, power-efficient circuit applications. Moreover, using superconducting interconnects can further reduce power consumption. Josephson junction field-effect transistors (JJFET, Fig. a) have recently emerged as a promising candidate for superconducting computing. JJFETs are particularly useful for low power consumption applications as they are operated with, in the superconducting regime, zero voltage drop across its source and drain. Feasibility of using JJFETs for logic operations has been explored [1]. Moreover, high noise margin has been demonstrated for the logic inputs [2]. Compared to single flux quantum approach, JJFET circuits can use designs similar to conventional silicon MOSFETs [1].For JJFETs to perform logic operations, the gain-factor (αR) value must be larger than 1. Here αR = dIc/d(Vg–Vt) x πΔ/Ic, Δ is the superconducting gap, Vg the gate bias voltage, Vt the threshold voltage. Ic ∝ exp(-L/ξc) is the critical supercurrent, where L is the channel length and ξc the carrier coherence length [3]. In a conventional JJFET, ξc ∝ (Vg–Vt)0.5 (Fig. b), and thus dIc/d(Vg–Vt) is small (Fig. c). This translates to a requirement of superconducting transition temperature of ~ 400K for αR larger than 1, far exceeding any recorded critical temperatures. As such, it is impossible to use conventional JJFETs for logic operations.Here, we propose a novel type of JJFET based on quantum phase transition, such as the excitonic insulator (EI) transition in a type-II InAs/GaSb heterostructure [4,5], for low-energy, power-efficient logic applications. The nature of the collective phenomenon in the EI quantum phase transition can provide a sharp transition of the supercurrent states (e.g., ξc ∝ (Vg–Vt)5 and dIc/d(Vg–Vt) very large, as shown in Figs. b and c, respectively) which will enable αR larger than 1 with an easy-to-achieve superconducting transition temperature of ~ 40K. In this talk, we will present some preliminary results demonstrating that indeed the gain factor in these quantum enhanced JJFETs can be greatly improved, thus making them a promising candidate for logic applications.Sandia National Laboratories is a multimission laboratory managed and operated by NTESS LLC, a wholly owned subsidiary of Honeywell International Inc. for the U.S. DOE’s NNSA under contract DE-NA0003525. This paper describes objective technical results and analysis. Any subjective views or opinions that might be expressed in the paper do not necessarily represent the views of the U.S. DOE or the United States Government.

Performance Analysis of Lake Water Cooling Coupled with a Waste Heat Recovery System in the Data Center

Data centers (DCs) require continuous cooling throughout the year and produce a large amount of low-grade waste heat. Free cooling and waste heat recovery techniques are promising approaches to reduce DC energy consumption. Although previous studies have explored diverse waste heat utilization strategies, there is a significant gap in combining waste heat recovery with lake water cooling in DCs. Therefore, this study proposed a system integrating lake water cooling with waste heat recovery for DCs. To evaluate the energy-saving performance of the suggested system, the influence of waste heat recovery locations and volumes has been investigated. An analysis of the improvement in system parameters is also conducted. The study’s findings highlight that targeted recovery of waste heat from sources like chilled water or air in server rooms can significantly reduce the cooling energy demand of the system. The results show that recovering heat from the return air of IT equipment can yield a remarkable power usage effectiveness (PUE) and coefficient of performance (COP) of 1.19 and 10.17, and the energy consumption of the cooling system is reduced to 10.06%. Moreover, the outcomes reveal the potential for substantial energy savings of up to 26.05% within the proposed system by setting the chilled water and air supply temperatures to 16 and 20 °C, respectively.

Methodology to Characterize Row Manifolds for High Power Direct to Chip Liquid Cooling Data Centers

Abstract Demand is growing for the dense and high-performing IT computing capacity to support artificial intelligence, deep learning, machine learning, autonomous cars, the Internet of things, etc. This led to an unprecedented growth in transistor density for high-end CPUs and GPUs, creating thermal design power (TDP) of even more than 700 watts for some of the NVIDIA existing GPUs. Cooling these high TDP chips with air cooling comes with a cost of the higher form factor of servers and noise produced by server fans close to the permissible limit. Direct-to-chip cold plate-based liquid cooling is highly efficient and becoming more reliable as the advancement in technology is taking place. Several components are used in the liquid-cooled data centers for the deployment of cold plate based direct to chip liquid cooling like cooling loops, rack manifolds, CDUs, row manifolds, quick disconnects, flow control valves, etc. Row manifolds used in liquid cooling are used to distribute secondary coolant to the rack manifolds. Characterizing these row manifolds to understand the pressure drops and flow distribution for better data center design and energy efficiency is important. In this paper, the methodology is developed to characterize the row manifolds. Water-based coolant Propylene glycol 25% was used as the coolant for the experiments and experiments were conducted at 21 °C coolant supply temperature. Two, six-port row manifolds' P-Q curves were generated, and the value of supply pressure and the flow rate were measured at each port. The results obtained from the experiments were validated by a technique called Flow Network Modeling (FNM). FNM technique uses the overall flow and thermal characteristics to represent the behavior of individual components.

Stochastic optimization for capacity configuration of data center microgrid thermal energy management equipment considering flexible resources

The rapid development of Internet and cloud computing technologies has led to the expansion of the scale of data centers, resulting in a continuous increase in the demand for data processing and storage. This not only results in higher energy consumption but also makes thermal management a key factor of energy management in data centers. A capacity configuration method is proposed for the core thermal management equipment of data center microgrids, electric boilers, cooling systems, and heat pumps. This method not only considers the uncertainty factors of the source and load but also effectively utilizes the characteristics of various flexible resources in data centers, such as the adjustable ability of different types of batch processing loads, air thermal inertia, and waste heat recovery, significantly improving the energy utilization efficiency and enhancing the system’s economy and flexibility. In addition, a wind power scenario generation method based on conditional least squares generative adversarial networks was proposed to improve the generation quality of wind power scenarios, and a random optimization model was constructed. Considering a certain province’s data center microgrid as a case study, the effectiveness of the model was verified, and a sensitivity analysis of the relevant parameters provided important references for microgrid planning. This study provides a new perspective and tool for the thermal management of data center microgrids and comprehensive utilization of renewable energy, which is of great significance for promoting the improvement of data center energy efficiency and sustainable development.

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.

Variational quantum circuit learning-enabled robust optimization for AI data center energy control and decarbonization

As the demand for artificial intelligence (AI) models and applications continues to grow, data centers that handle AI workloads are experiencing a rise in energy consumption and associated carbon footprint. This work proposes a variational quantum computing-based robust optimization (VQC-RO) framework for control and energy management in large-scale data centers to address the computational challenges and overcome limitations of conventional model-based and model-free strategies. The VQC-RO framework integrates variational quantum circuits (VQCs) with classical optimization to enable efficient and uncertainty-aware control of energy systems in AI data centers. Quantum algorithms executed on noisy intermediate-scale quantum (NISQ) devices are used for value function estimation trained with Q-learning, leading to the formulation of a robust optimization problem with uncertain coefficients. The quantum computing-based robust control strategy is designed to address uncertainties associated with weather conditions and renewable energy generation while optimizing energy consumption in AI data centers. This work also outlines the computational experiments conducted at various AI data center locations in the United States to analyze the reduction in power consumption and carbon emission levels associated with the proposed quantum computing-based robust control framework. This work contributes a novel approach to energy-efficient and sustainable data center operation, promising to reduce carbon emissions and energy consumption in large-scale data centers handling AI workloads by 9.8 % and 12.5 %, respectively.

Renewable Energy in Data Centers: The Dilemma of Electrical Grid Dependency and Autonomy Costs

Renewable Energy in Data Centers: The Dilemma of Electrical Grid Dependency and Autonomy Costs

Optimizing Data Centre Energy Efficiency with Dynamic Resource Allocation and Intelligent Cooling Management through Machine Learning

Our research aims to improve energy efficiency in data centres by combining cloud computing infrastructure with machine learning techniques. We propose that dynamic resource assignment, combined with intelligent optimisation of cooling systems, can reduce power waste and operational costs. Real-time sensor data from various data centre components such as servers, cooling systems, and power distribution units is collected and fed into machine learning models for analysis. In this way, we can create power arrangements that are tailored to the resources and needs of various applications. The experimental results show that energy consumption has been reduced by an average of 30% compared to traditional methods. Furthermore, our machine learning models are quite accurate in predicting cooling system performance. For example, an Artificial Neural Network (ANN) has an accuracy rate of 98.78%. This result demonstrates the efficacy of our approach in promoting energy efficiency and operational performance in data centres: it not only provides a scalable, cost-effective solution to industry energy efficiency challenges, but it also improves day-to-day data centre operations by reducing electrical consumption. Our approach, which is based on dynamic allocation of computational resources and real-time data analysis for optimising cooling systems, not only saves energy but also improves operational efficiency. With energy in mind, we must all work towards a more sustainable and green approach to data centre management. Future research can look into other potential optimisations, as well as issues with scalability and application in real-world data centre environments.

A systematic review of green-aware management techniques for sustainable data center

Cloud computing is one of the powerful engines driving global industrial upgrading and the booming digital economy. However, the explosive growth of cloud data centers (DCs) has resulted in inevitable energy consumption and carbon emission problems. Therefore, constructing energy-efficient and sustainable DCs will be essential for green cloud computing. This review makes several efforts to thoroughly investigate and track the research progress and routes to sustainable DCs. Firstly, we construct a new conceptual model of sustainable DCs to cover cutting-edge research results and indicate future evolutionary directions. Secondly, this review provides a comprehensive survey of five topics from a technical perspective: workload management, virtual resource management, energy management, thermal management, and waste heat recovery. Subsequently, some real-world datasets relevant to the topics, including workload traces, renewable energy data, and electricity price traces, have been specifically collected to support researchers in conducting further research. Finally, based on observations of existing works, we highlight some salient technical challenges and promising solutions to provide sensible energy and carbon reduction suggestions in sustainable DCs.

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