Power Usage Effectiveness Research Articles (Page 1)

The increase in data center facilities has led to higher energy consumption and a larger carbon footprint, prompting improvements in thermal environments for energy efficiency and server lifespan. Existing literature studies often overlook categorizing equipment for power usage effectiveness (PUE), addressing power efficiency measurement limitations and employee thermal comfort. These issues are addressed through an investigation of the PUE metric, a comparative analysis of various data center types and their respective cooling conditions, an evaluation of PUE in relation to established thermal standards and an assessment of employee thermal comfort based on defined criteria. Thirty-nine papers and ten websites were reviewed. The results indicated an average information technology (IT) power usage of 44.8% and a PUE of 2.23, which reflects average efficiency, while passive cooling was found to be more applicable to larger-scale data centers, such as Hyperscale or Colocation facilities. Additionally, indoor air temperatures averaged 16.5 °C with 19% relative humidity, remaining within the allowable range defined by ASHRAE standards, although employee thermal comfort remains an underexplored area in existing data center research. These findings highlight the necessity for clearer standards on power metrics, comprehensive thermal guidelines and the exploration of alternative methods for power metrics and cooling solutions.

In dynamic environments like cloud computing, the internet of things (IoT), and cyber-physical systems, where conventional rule-based adaptation mechanisms frequently fall short of maintaining optimal performance in the face of uncertainty and change, self-adaptive software systems (SASS) are becoming more and more important. A promising remedy that allows systems to learn and adapt on their own through trial-and-error interactions is Reinforcement Learning (RL). After a thorough screening of 1,248 papers, 68 quantitative studies were chosen for analysis in this systematic review, which examines developments in RL for dynamic optimisation of SASS from 2021 to early 2025. Value-based, policy gradient, multi-agent, and hybrid/meta-learning approaches are the main RL methodologies identified in the review, which also looks at how they are applied in fields like cybersecurity, cloud resource management, and autonomous systems. The findings indicate that Cloud systems reduced average cost by 34.7% using PPO-based solutions, cybersecurity systems improved attack detection speed by 22.1% and false positive rates by 18.3%, and autonomous systems reduced energy consumption by 40% and adaptation latency by 27.5% in IoT and swarm robotics. Policy gradient methods (41%) dominate continuous control tasks, with PPO used in 27% of studies. Value-based approaches (32%) dominate discrete action domains, with deep Q-networks (DQN) variants used in 78% of cloud resource allocation studies. Multi-agent RL accounts for 18% of studies, with Multi-agent deep deterministic policy gradient (MADDPG) - 62% and QMIX (38%) being the most used. Serverless computing cut cold-start times by 35%, data centre optimisation lowered power usage effectiveness (PUE) by 15%, and RL-driven intrusion detection systems identified zero-day threats with 92% accuracy. Reward design difficulties were found in 63% of experiments, sample inefficiency required 1.2M episodes to converge, and real-world multi-agent reinforcement learning (MARL) deployments performed 23% worse than models. Metal analysis effect resulted in 95% in cost reduction, latency improvement and adaptation speed respectively. Practical adoption is limited because so few studies use standardised benchmarks or address safety and interpretability. In order to close the gap between research and practical implementation, the article ends by outlining open research questions and promoting formal verification, transfer learning, and hybrid learning.

Data centers contribute to roughly 1% of global energy consumption and 0.3% of worldwide carbon dioxide emissions. The cooling system alone constitutes a substantial 50% of total energy consumption for data centers. Lowering Power Usage Effectiveness (PUE) of data center cooling systems from 2.2 to 1.4, or even below, is one of the critical issues in this thermal management area. In this work, a digital twin system of an Intelligent Data Center (IDC) prototype is designed to be capable of real-time monitoring the temperature distribution. Moreover, aiming to lower PUE, Deep Q-Learning Network (DQN) is further established to make optimization decisions of thermal management during cooling down of the local hotspot. The entire process of thermal management for IDC can be real-time visualized in Unity, forming the virtual entity of data center prototype, which provides an intelligent solution for sustainable data center operation.

Green computing becomes another important aspect of sustainable technology from the standpoint of energy usage management as the provision of energy-efficient functionality of data centers is essential to deal with environmental issues. In this study, we propose an evolution of energy efficient data centers via AI-based integration. Adopting a use-case driven methodology, this study creates and tests AI modularized models that respond to dynamically forecast workloads, predict energy demand loads, and optimize active cooling systems. With the help of machine learning algorithms, real-time monitoring and predictive analytics, this methodology can identify opportunities for energy blanketing, allowing for a significant decrease in energy consumption without affecting performance. Multi data center experiments showing a measured PUE (power usage effectiveness) and Carbon reduction An AI-driven dynamic cooling strategy reduced energy consumption by as much as 25% and intelligent workload distribution improved system efficiency by 15%. These results highlight the ability of AI to dramatically reduce the environmental impact of high energy computing systems, helping to ensure sustainable growth of data center operations in the future. Finally, this paper ends with the broader ramifications where green computing advances across multiple sectors like the energy-intensive AI, advocating similar efforts for other energy-intensive industries in line with global sustainability objectives. This also highlights the importance of collaboration between academia and industry to advance cleaner computing technologies more broadly.

Data centres require energy to power their computing equipment as well as to maintain proper environmental conditions through their extensive cooling systems. Data centres are a key part of digital infrastructure, but use a lot of energy, especially for cooling and computing. This paper explores energy trends and reviews solutions like CRAC systems, chilled water cooling, and free cooling. It also discusses energy efficiency using measures like Power Usage Effectiveness (PUE) and shows real examples of improvement. The goal is to help design greener, more efficient data centres. The research investigates the main elements that determine energy consumption in data centres. The paper examines two emerging technologies and strategies for decreasing energy usage, which include CRAC systems and free cooling, and chilled water systems. It is estimated that annually in the U.S, the data centres consume roughly 50% of electricity, mainly by the equipment. The cooling needs for Heating, Ventilation, and Air Conditioning (HVAC) are estimated to be up to 40% using computer room air-conditioners to cool down the equipment, such as servers, and other IT equipment in the data centres. This paper discusses the high energy needs of data centres and also reduces energy use by making systems as efficient as possible. Providing sustainable data centres is the energy goal so as to maximise energy from renewable systems. Data centres need a comprehensive strategy that combines operational excellence with environmental responsibility and financial sustainability to enhance their energy efficiency. Future research must create regionally adaptable solutions that reduce data centre environmental impact because digital performance expectations will continue to grow.

Virtual machine (VM) placement in cloud datacenters is a complex multi-objective challenge involving trade-offs among energy efficiency, carbon emissions, and network performance. This paper proposes NCRA-DP-ACO (Network-, Cost-, and Renewable-Aware Ant Colony Optimization with Dynamic Power Usage Effectiveness (PUE)), a bio-inspired metaheuristic that optimizes VM placement across geographically distributed datacenters. The approach integrates real-time solar energy availability, dynamic PUE modeling, and multi-criteria decision-making to enable environmentally and cost-efficient resource allocation. The experimental results show that NCRA-DP-ACO reduces power consumption by 13.7%, carbon emissions by 6.9%, and live VM migrations by 48.2% compared to state-of-the-art methods while maintaining Service Level Agreement (SLA) compliance. These results indicate the algorithm’s potential to support more environmentally and cost-efficient cloud management across dynamic infrastructure scenarios.

Efficient cooling and heat recovery systems are becoming increasingly critical in large-scale commercial and industrial facilities, especially with the rising demand for sustainable energy solutions. Traditional air-conditioning and refrigeration systems often dissipate significant amounts of waste heat, which remains underutilized. This study addresses the challenge of harnessing low-potential waste heat from such systems to support fifth-generation district heating and cooling (5GDHC) networks, particularly in moderate-temperate regions like Flanders, Belgium. To evaluate the technical and economic feasibility of waste heat recovery, a methodology is developed that integrates established performance metrics—such as the energy efficiency ratio (EER), power usage effectiveness (PUE), and specific cooling demand (kW/t)—with capital (CapEx) and operational expenditure (OpEx) assessments. Empirical correlations, including regression analysis based on manufacturer data and operational case studies, are used to estimate equipment sizing and system performance across three operational modes. The study includes detailed modeling of data centers, cold storage facilities, and large supermarkets, taking into account climatic conditions, load factors, and thermal capacities. Results indicate that average cooling loads typically reach 58% of peak demand, with seasonal coefficient of performance (SCOP) values ranging from 6.1 to a maximum of 10.3. Waste heat recovery potential varies significantly across building types, with conversion rates from 33% to 68%, averaging at 59%. In data centers using water-to-water heat pumps, energy production reaches 10.1 GWh/year in heat pump mode and 8.6 GWh/year in heat exchanger mode. Despite variations in system complexity and building characteristics, OpEx and CapEx values converge closely (within 2.5%), demonstrating a well-balanced configuration. Simulations also confirm that large buildings operating above a 55% capacity factor provide the most favorable conditions for integrating waste heat into 5GDHC systems. In conclusion, the proposed approach enables the scalable and efficient integration of low-grade waste heat into district energy networks. While climatic and technical constraints exist, especially concerning temperature thresholds and equipment design, the results show strong potential for energy savings up to 40% in well-optimized systems. This highlights the viability of retrofitting large-scale cooling systems for dual-purpose operation, offering both environmental and economic benefits.

To improve the energy economic efficiency of Data Centers (DCs) in steel enterprises, a centralized heating scheme for waste heat recovery based on the Co-ah cycle is proposed. This scheme establishes a thermoelectric connection between the self-owned power plant of the steel enterprise and the DC, creating a waste heat recovery centralized heating system for a 15 MW DC. The energy efficiency indicators, environmental benefits, economic feasibility, and adaptability of the system are evaluated. The results show that the system can effectively recover waste heat from the DC, significantly reducing cooling electricity consumption during the heating season and decreasing original heating steam consumption by about 25%. Compared to DC using free cooling, the annual operating cost is reduced by 9.7%, with a dynamic payback period for equipment of 6–7 years. The system saves 3,671.5 tons of standard coal and reduces CO2 emissions by 1,615 tons annually compared to DC using isolated free cooling and traditional heating systems. The Improved Power Usage Effectiveness (PUE’) of the system is 1.195, and the Energy Reuse Effectiveness (ERE) is 0.769, outperforming the free cooling’s index of 1.341, although Exergy Reuse Effectiveness (ExRE) is slightly higher than that of free cooling. This system offers mutual benefits for self-owned power plant, DC, and heating companies, achieving a win-win operational state through suitable energy trading prices. The research conclusion provides valuable reference for the future investment and operation of DCs in steel enterprises.

To promote the coordinated development of green data centers, the “Channel Computing Resources from East to West (CCR-EW)” project plans to migrate computing power demand from the eastern region to China’s national computing hubs in the west. Considering the many factors that affect the migration of computing loads between data centers, there is an urgent need to develop an appropriate migration strategy to facilitate the CCR-EW project and ensure service reliability. Therefore, this paper proposes a load migration strategy between data centers and a load migration strategy in a data center taking carbon emission and operating cost as the optimization objectives. The primary constraints, namely power usage effectiveness (PUE), electricity price, and carbon emission intensity are duly accounted for as the main constraints. To this end, the optimization model is constructed for these two strategies, considering workload, delay constraints, by leveraging queueing theory to enhance the optimization of renewable energy utilization. Finally, the optimization model is applied to China’s national computing hubs. The application results show that these models reduce the carbon emission of the data center, improve service reliability, promote the decision of computing load migration, and enhance the efficiency of renewable energy utilization.

Heat transfer dependence of power usage effectiveness of an augmented two-phase immersion cooling system for high-power servers

This paper introduces two distinct approaches for efficiently cooling high-density data centers with a capacity of up to 1168 kW, aiming to achieve the optimal cooling method for extracting heat from electronic servers. The investigation involves a comparative analysis between air-based cooling, incorporating modifications to hot aisle containment, and liquid-based cooling utilizing a two-phase immersion system. Three-dimensional simulations are conducted employing computational fluid dynamics (CFD) analysis. Furthermore, the simulation examines the influence of introducing copper coating materials to server components in liquid cooling, varying the thickness to enhance heat transfer efficiency. The analysis demonstrates that, in liquid cooling, the maximum server temperature diminishes to 65 °C, and power usage effectiveness improves from 2 in the air cooling system to 1.01 in liquid cooling. The application of copper coating results in a 13% enhancement in the gradual temperature distribution among servers, especially in high-power servers, and eliminates most hot spots by decreasing the maximum temperature. However, the simulation indicates a comparatively modest temperature improvement of 8% when coating low-power servers.

This paper introduces the SecureSustainNet Framework, a novel approach designed to enhance security and sustainability within data centers. The framework employs a multi-objective optimization technique that harmonizes security measures with energy efficiency. It includes six core algorithms: Intrusion Detection System with Anomaly Detection, AES-256 Encryption with Optimized Key Management, Role-Based Access Control with Dynamic Policy Adjustment, Dynamic Resource Allocation, Energy Consumption Monitoring and Optimization, and Renewable Energy Integration. The framework was implemented and evaluated using the ns-3 network simulator. The evaluation results demonstrate significant advancements in both security and sustainability. The Intrusion Detection System achieved an Anomaly Detection Rate (ADR) of 98.7%, reflecting high accuracy in threat identification. Encryption overhead was minimized to a 2.5% increase in processing time, showcasing efficient performance. The Role-Based Access Control system attained an effectiveness of 97.4% in preventing unauthorized access. Resource Utilization Efficiency (RUE) reached 85.2%, indicating effective resource management. The framework also achieved a 15.6% reduction in energy consumption and a Power Usage Effectiveness (PUE) of 1.25, signifying improved energy efficiency. These results underline the SecureSustainNet Framework’s effectiveness in integrating robust security measures with advanced sustainability practices, presenting it as a valuable model for optimizing data center operations.

With the rapid increase in global data demand, data centers have emerged as the primary drivers of digital transformation and the foremost contributors to worldwide information and communication technology (ICT) energy consumption. This article examines the twin necessity of constructing environmentally sustainable and secure data centers through the integration of energy-efficient designs with sophisticated cybersecurity measures. This research develops a cohesive AI-driven framework by synthesizing eleven contemporary studies published between 2023 and 2024, including contributions from Kaur et al., Hasan et al., Mahmud et al., Rahman et al., and Faisal et al., which integrates sustainability metrics (e.g., Power Usage Effectiveness and carbon intensity) with cyber-resilience indicators (e.g., anomaly-detection accuracy and mean-time-to-respond). The research delineates three essential integration layers: (a) sustainable computing and intelligent resource administration, (b) AI-augmented cybersecurity utilizing big data and blockchain technology, and (c) governance via management information systems (MIS). Findings indicate that the proposed AI–Green Secure Data Center Framework can decrease energy usage by 20–25% and enhance threat-response efficiency by 30–40%. The framework promotes a novel paradigm for sustainable and resilient digital infrastructure in the context of Industry 5.0 by integrating ecological stewardship with digital security.

Parameters of Performance: A Deep Dive into Liquid-to-Air CDU Assessment

Datacenters are becoming one of the most significant worldwide consumers of electricity and sources of carbon emissions. With the end of Dennard's scaling, and as cloud datacenter power usage effectiveness becomes close to optimal, we can no longer rely on hardware advancements alone to sustainably meet growing computational needs. System software must play a bigger role in optimizing energy consumption of applications. We argue the operating system (OS), at the heart of datacenter resource allocation and scheduling decisions, must be made energy-aware. This paper has two goals. First, we show how Linux can be made energy-aware without making any kernel changes, by introducing a new energy accounting framework, Wattmeter. Wattmeter uses eBPF functions to efficiently measure per-process energy consumption at millisecond-scale granularity with low overhead. Second, we show how this information can be used to make energy-informed scheduling decisions, with two proof-of-concept scheduling policies: a policy that equalizes energy across processes, and one that caps the amount of energy that can be consumed by a process. This paper represents a first step in making the operating system energy-aware, and demonstrating how that capability can be used to control applications' energy consumption.

This article systematically analyzes the transformative impact of Artificial Intelligence (AI) and Machine Learning (ML) technologies on modern data center operations. Through an extensive review of implemented cases and empirical data from multiple data centers, the article demonstrates how AI-driven solutions significantly enhance operational efficiency, reduce maintenance costs, and improve infrastructure reliability. The findings indicate that predictive maintenance algorithms achieve a 47% reduction in unexpected equipment failures, while ML-based resource optimization leads to a 31% improvement in resource utilization rates. The article examines integrating deep learning models for real-time energy management, resulting in an average 23% reduction in cooling costs and a 0.15 improvement in Power Usage Effectiveness (PUE). Additionally, the article analyzes the implementation of AI-powered security frameworks, which demonstrated a 92% accuracy rate in anomaly detection and reduced false positives by 76% compared to traditional rule-based systems. The article also presents a novel framework for capacity planning using neural networks, achieving an 89% accuracy in demand forecasting over 12 months. These findings provide valuable insights for data center operators and establish best practices for implementing AI/ML solutions in mission-critical infrastructure environments. The article concludes with recommendations for overcoming integration challenges and a roadmap for future technological adoption.

Indirect evaporative cooling (IEC), which utilizes natural cooling sources, is an advanced and promising technology to reduce the energy consumption of cooling systems in a data center (DC). This study presents a model of an IEC air-conditioning unit in a DC using TRNSYS simulation software validated using actual operational data to investigate the adaptability of IEC units in data centers located in regions with varying humidity levels, providing a reference for their application and promotion in DCs. Based on this premise, the authors analyzed the meteorological characteristics of Urumqi (a dry region), Beijing (a region with medium humidity), and Shanghai (a region with high humidity), which are representative cities in different humidity zones. The analysis identified the annual operating hours of the unit’s three operation modes, including fresh-air indirect heat transfer (FAIHT), IEC, and hybrid. Simultaneously, the authors conducted a simulation of the unit’s yearly energy consumption and determined time change curves for annual energy consumption, hourly coefficient of performance (COP) throughout the year, and mechanical cooling in various locations. The results indicate that IEC air-conditioning systems are highly effective in promoting the efficiency of data centers in various humidity regions. Dry locations demonstrate the greatest adaptability, followed by regions with medium humidity and, finally, regions with high humidity. The findings indicate that IEC units provide significant energy efficiency and cost-effectiveness when deployed in typical urban DCs across various humidity zones in China. The average annual power-usage effectiveness (PUE) of each city’s DC utilizing the unit is less than 1.3, and the unit’s annual operational cost savings exceed 30%.

Performance and energy consumption study of a dual-evaporator loop heat pipe for chip-level cooling

This paper presents a comprehensive quantitative model for quantitative assessment of the lifecycle costs and environmental impacts of computing infrastructure, with a focus on internet data centers (IDCs) and high-performance computing (HPC) facilities. The key innovation lies in the integration of interdisciplinary cost evaluation and carbon emission methods for the establishment of this quantitative model. This framework, which outlines key cost components and carbon emission factors, enables the calculation of total costs, electricity expenses, and greenhouse gas emissions throughout the lifecycle of infrastructure. With IDCs as a case study, the research clarifies the intricate cost structure associated with equipment procurement, energy usage, land acquisition, and operational expenses. This paper provides an in-depth understanding of the cost structure and environmental impact of computing infrastructure in support of sustainable decision-making in its development.•Based on established cost estimation methods, such as Lifecycle Cost Analysis (LCCA) and the Analogous Estimating Method, this study examines costs across construction and operation phases.•The Emission Factor Method is used to quantify environmental impact, emphasizing the significance of regional energy mix and power usage effectiveness (PUE).

A model predictive control for a multi-chiller system in data center considering whole system energy conservation