Search

• Equation-based object-oriented Modelica models for data center cooling systems are built and released in the open-source Modelica Buildings library. • Together with the Electrical package in the Modelica Buildings library, the introduced package can simulate the thermal dynamics as well as the electrical dynamics for conventional and renewable data centers. • New strategies toward emergency operations for the data center cooled by chillers and waterside economizers are investigated. In this paper, we introduce a newly developed open source data center package in the Modelica Buildings library to support modeling and simulation of cooling and control systems of data centers. The data center package contains major thermal and control component models, such as Computer Room Air Handler, Computer Room Air Conditioner, models of different subsystem configurations such as chillers with differently configured waterside economizers, as well as templates for different systems. Two case studies based on this package are performed to investigate the performances of the cooling and electrical system under normal conditions and emergency situations such as a blackout: one is for a data center powered by conventional energy, and the other is for a data center powered by renewable energy. Simulation results show that the dynamic modeling and multi-domain simulation in the Modelica-based tool make it convenient for users to investigate both normal and emergent operations for conventional and renewable data centers. [ABSTRACT FROM AUTHOR]

[Display omitted] • Definition and indicators of the local hotspots are clarified. • The main problems of weak airflow distribution are deeply analyzed. • Thermal management technologies are summarized by different scales for guidance. • Current limitations and future trend of hotspots elimination methods are discussed. The data centers consume a significant amount of energy, with cooling demand accounting for roughly 40% of the total energy usage. The primary reason for this is the presence of the local hotspots. To eliminate these hotspots, the standard approach is to lower the set point temperature of the whole computer room air conditioner (CRAC), resulting in overcooling and energy waste. This study provides an overview of the recent research on local hotspot management strategies, with a focus on the overheated servers as the most likely location of hotspots. To quantify the hotspots, this study provides a clear definition and summarizes the indicators from different aspects. It analyzes the correlation between the inlet and outlet air temperature and the overheating issues, which is often caused by high server loads or poor airflow distribution. And the most fundamental manifestation of the hotspots is hot air recirculation. This study analyzes potential solutions, such as optimizing airflow distribution and shortening the cold airflow path through design, control, and application. Additionally, the study discusses the limitations of the recent hotspots elimination research and proposes future perspectives. The aim of this review is to provide helpful knowledge on local hotspot elimination in the data centers, with consideration for integrated thermal and airflow distribution management strategies. The conclusion may assist in designing, operating, and retrofitting energy-efficient data centers in practice. [ABSTRACT FROM AUTHOR]

In recent years, water-side free cooling has been broadly applied in data centers to lessen the high cost of energy consumption. The energy-saving opportunities of free cooling have been recognized by numerous studies across the world. In this paper, the field test and energy performance evaluation were performed in an operational data center with water-side free cooling, located in Chongqing, China. Existing problems were found during tests, including the insufficient cooling capacity of towers in wintertime and the inability to activate partial-free cooling. The results of the tests show that the reduction in Power Usage Effectiveness (PUE) is much lower than the expected. To understand the marked discrepancies, the performance of the cooling tower was discussed in terms of the approach temperature and actual cooling capacity. The maximum cooling capacity in winter is only about 50 % of the rated. With the consideration of the cooling tower configuration performance in different climate conditions, the conservative energy effects were calculated in 34 provincial cities of China. To prolong the hours available, optimized operation schemes of cooling towers are presented. The available hours for free cooling could be increased from 730 h to 1300 h when operating at 67 % of rated flow rate, resulting in a decrease in the energy saving of chillers between 8.44 % and 21.05 %. [ABSTRACT FROM AUTHOR]

Row-based cooling has been recommended for cooling scenarios with high power densities. Existing studies on row-based cooling have mostly focused on good results, and the effectiveness and limitations of it need to be evaluated to optimize it. In this study, some airflow problems were observed by a field experiment for a data center with row-based cooling. Subsequently, a simulation corresponding to this data center was performed to fully learn and verify some airflow conditions below those expected, such as hot spots in some racks and vortex airflow, where the Supply Heat Index of modules 1 and 2 are 0.37 and 0.31. To further determine the airflow characteristics of row-based cooling, a comparative simulation of airflow distribution of row-based cooling and raised-floor cooling was performed. Finally, the effectiveness and limitations of row-based cooling were presented. The results showed that row-based cooling only decreases the Return Temperature Index by 0.4%. The limitations of it include vortex airflow, hot spots in head racks and adjacent racks of in-row coolers, and interaction between different modules; they are easy to occur in row-based cooling and some improvement suggestions are presented in future work. [ABSTRACT FROM AUTHOR]

Highlights • Equation-based Modelica is utilized to identify and evaluate operational issues for a data center cooling system. • Sequential Search Technique is used to find the most cost-effective retrofit solution. • The best energy efficiency measure together with an optimized supply air temperature setpoint can save up to 24% of cooling energy in the data center. Abstract Data center cooling accounts for about 1% of electricity usage in the United States. Computer models are pivotal in designing and operating energy-efficient cooling systems. Compared to conventional building performance simulation programs, the equation-based object-oriented modeling language Modelica is an emerging approach that can enable fast prototyping and dynamic simulation of cooling systems. In this case study, we first modeled the cooling and control systems of an actual data center located in Massachusetts using the open-source Modelica Buildings library, and then calibrated a baseline model based on measurement data. The simulation of the baseline model identified several operation-related issues in the cooling and control systems, such as degraded cooling coils, improper dead band in control settings, and simultaneous cooling and heating in air handlers. Afterwards, we used a sequential search technique as well as an optimization scheme to investigate the energy saving potentials for different energy efficiency measures aiming to address the abovementioned issues. Simulation results show potential energy savings up to 24% by resolving identified control-related issues and optimizing the supply air temperature. [ABSTRACT FROM AUTHOR]

Highlights • The art on airflow distribution and management in data center is reviewed. • Experimental and numerical investigations of several typical cases are summed. • Energy efficiency of air distribution must be based on the reliability of data center. • HACS is a promising air management system due to its flexibility and extendability. • Improvement methods of vertical airflow uniformity needs to be developed. Abstract Rapid expansions of high-density data centers have been triggered worldwide owing to the high demand of cloud computing over the past decade. The airflow distribution and management have great association with the thermal performance of the typical underfloor air distribution (UFAD) data center. Therefore, numbers of configurations and methods have been proposed to improve the thermal performance of data centers with various operation conditions. This paper reviews the ventilation configurations of data center and presents the geometric effects of underfloor plenum on airflow. The experimental and numerical investigations on typical cases as well as the relevant parameters are also summed up. Various air management methods including cold aisle containment system(CACS), hot aisle containment system(HACS) and vertical exhausted duct system(VEDS) are analyzed on the benefits and limitations. Based on the enormous amount of research on the traditional UFAD system with hot aisle/cold aisle (HA/CA) arrangement, it can be concluded that the practices of data centers have a tendency of being more flexible and extendable, while the literature on the corresponding cases which applied HACS or other air distribution systems remains scarce. Moreover, beside the improvement of the airflow uniformity along the cold aisle, methods for enhancing the airflow uniformity along the height direction are also encouraged to be developed. [ABSTRACT FROM AUTHOR]

• Cooling water operation parameters under different heat sink were optimized. • Fitting correlation on affordable extreme atmospheric temperature is given. • The B-type system can achieve significant energy savings compared to the A-type system. • The B-type can meet the year-round cooling demand via a cooling tower cold source. • Optimal water flow rate fluctuation clearly responds to different atmospheric conditions. Data center, as a special building form, has attracted much attention because of its increasing energy consumption. In order to reduce the energy consumption of data center caused by cooling system, research on the water-cooled heat sinks used to cool chips has become increasingly important with the steep increase in the heat flux density of data centers. To explore the effect of different water-cooled heat sinks on cooling system performance, a server water-cooled system for the server cabinet was built, and the chip heat dissipation performance, optimal thermal management, and system energy consumption were simulated by equipping a cooling tower and two types of fin-type water-cooled heat sinks (A-type and B-type). By comparing the effects of A- and B-type heat sinks on the server water-cooled system performance, the optimal water working condition was obtained under different atmospheric conditions by minimizing the system power consumption to guarantee that the chip temperature is no higher than 70 °C. It was found that the optimal cooling water flow rate fluctuates in a large range under different atmospheric conditions. A server water-cooled system equipped with B-type can achieve significant energy savings compared to those from A-type system when the atmospheric relative humidity and temperature are high. Considering a case study from Tianjin, China, it is shown that the month-averaged power usage effectiveness in June is 1.2 for A-type and 1.14 for B-type systems. [ABSTRACT FROM AUTHOR]

• The finned tube heat exchangers with ultra-low temperature surfaces, a novel passive cooling system, directly cool down the data center. • The effect of ultra-low temperature surfaces on the temperature distribution in the room is analyzed through experiments. • The TSD value evaluates the degree of temperature stratification. As artificial intelligence has been leaping forward, data centers (DCs) are only capable of meeting to the daily needs of people through uninterrupted operations. A power outage or cooling system failure can result in high temperatures, trigger data center shutdowns, and even damage IT equipment. In terms of several special DCs, a prolonged cooling system interruption may lead to catastrophic and incalculable consequences. Accordingly, a novel passive cooling system was proposed in this study to address the long-term interruption of air conditioning systems. Liquid nitrogen (LN 2) was driven by the pressure of the cryogenic liquid container and flowed into finned tube heat exchangers (FTHEs) in the DC. LN 2 , i.e., a refrigerant, can absorb heat through vaporization and give FTHEs ultra-low temperature surfaces to drop the indoor temperature. The effects of refrigerant flow rate, temperature, and extreme indoor initial temperature on the indoor temperature distribution were examined. As indicated by the results, increasing the flow rate and decreasing the refrigerant temperature increased the room cooling rate while exacerbating the degree of temperature stratification (TSD). Compared with 35 kg/h, the cooling effect was enhanced by 1.5 times for a flow rate of 50 kg/h, whereas the TSD value near FTHEs was nearly 14 °C/m, approximately twice as high as that of 35 kg/h. Slightly raising the initial ambient temperature led to the rise of the cooling effect, whereas this method slightly affected the temperature stratification. This study provides a novel idea for solving long-term DC cooling system failure while laying a basis for the effect of ultra-low temperature surfaces on ambient temperature distribution. [ABSTRACT FROM AUTHOR]

• Circulating air is treated with dual-temperature chilled water in two stages. • Multiple natural energies are used to produce dual-temperature chilled water. • Longer free cooling duration and lower energy consumption are realized. • Dual-temperature chilled water system has very short payback period. Indoor high-temperature circulating air is only treated with one type of low-temperature chilled water in a conventional waterside free cooling system, resulting in a short free cooling duration and high energy consumption in a data center. This study develops an air conditioning system using dual-temperature chilled water for a two-stage treatment of air. Mathematical models are developed to calculate the system performance throughout the year. A data center located in Hohhot, China is taken as a case, and the energy consumption, energy efficiency, and economics of the grading treatment system are analyzed and compared with those of two conventional systems (i.e., waterside economizer and air-cooled heat exchanger systems). The results show that the grading treatment system achieves increased free cooling duration and energy efficiency in the data center compared with the conventional systems. Compared with the conventional waterside economizer and air-cooled heat exchanger systems, the grading treatment system realizes annual energy-saving rates of 27.4% and 25.4%, energy efficiency increments of 37.8% and 34.0%, and payback periods of 2.55 and 2.59 years, respectively. [ABSTRACT FROM AUTHOR]

[Display omitted] • Generative adversarial network is proposed to generate the training dataset for the building energy system. • The imbalanced dataset issue of FDD model is solved through GAN-based adversarial learning to avoid missing alarms. • An incremental learning SVM using optimal 8 features is developed to diagnose the refrigerant leakage faults in data center HVAC system. • Offline model updating strategy is developed and validated using the experimental datasets. The automated fault diagnosis and smart management of heating, ventilation and air conditioning (HVAC) system is essential to the reliability of data centers. The machine learning is an efficient way to develop the smart management tools for HVAC system using historical data. In real data centers, however, the quantity of fault-free data is much more than that of fault data. As a result, the imbalanced training data limits the diagnosis capacity of machine learning models. The deep learning-based generative adversarial network is proposed to integrate with an incremental learning SVM model to diagnose the commonly occurred faults of data center air conditioning system. The adversarial learning between generator and discriminator generates the data of minority class for training purpose in HVAC system. The 8 sensitive features are selected as the inputs through three-step optimal selection strategy: manual screening, correlation feature selection and redundancy feature removing. The incremental learning strategy is proposed to update the FDD model regularly. The refrigerant leakage faults with intensities of 10%, 20%, 30% and 40% are tested and validated under various operational conditions. The experimental results show that the incremental learning SVM integrated with deep learning GAN reaches the acceptable diagnosis accuracies. [ABSTRACT FROM AUTHOR]

• Actual LCI data collected for designed climate-friendly data center cooling device. • Nine different scenarios compared for the manufacturing process. • Carbon intensity between 39565,135 – 40071,702 CO 2 eq for the scenarios. • Water intensity between 40,0570–43,1801 m3 for the scenarios. • Renewable electricity use and waste management are critical for the manufacturing. The main aim of this study is to evaluate the manufacturing process's carbon footprint, water footprint, and environmental impacts for a newly designed cooling device in a data center. The results were compared for the different conditions of the manufacturing process: Use of wind electricity, use of solar electricity, use of electricity from the interconnected system, different waste amounts, and different forklifts. According to the characterization results with the Recipe (2016) method, wind electricity has less environmental impact in most of the analyzed categories except mineral resource scarcity. The scenarios with waste management offer the best environmental results for mineral resource scarcity. Normalization results define human health effects as the dominating category in which wind electricity use provides the best environmental results. Based on the IPCC 2013 GWP 100a results, the carbon footprint values for the scenarios varied between 39565,135–40071,702 CO 2 eq. The base scenario has the highest carbon footprint. Based on the Water Scarcity method, the water footprint varied between 40,0570–43,1801 m3. The scenarios with the highest water footprint are the base scenario and the scenarios where forklift replacement is made. The lowest water footprint values were achieved with the use of wind electricity. It has been determined that replacing a forklift does not contribute much to reducing carbon footprint, water footprint, or other environmental impacts. On the contrary, using renewable electricity and waste reduction significantly contribute to reducing the environmental effects. This study shows the importance of renewable electricity use and waste management for a newly designed device manufacturer. Therefore, a life cycle assessment study is beneficial for choosing the manufacturing path for a newly developed device. [ABSTRACT FROM AUTHOR]

It is important to cool data centers for improving IT equipment reliability through indoor temperature control. The air conditioning energy consumption is almost 38% of the total energy consumption. Free cooling in cool days is a promising technology for energy conservation. A pump-driven loop heat pipe (PLHP) system for free cooling was developed and installed in a data center project, which has a high energy efficiency ratio (EER) compared to air conditioners. The running performance of the PLHP system in the case project was tested and studied. The fitting rule of EER and outdoor temperature was obtained. The hourly energy savings and payback periods of the PLHP system for different cities of different climate regions in China were analyzed basing on the fitting formula compared to air conditioners. A national distribution map was developed to show the annual energy saving ratio in China. Results show that the PLHP system provides remarkable energy savings and the indoor temperature is remained at 18°C to 25°C. About 74.2% of Chinese cities lie in regions suitable for applications, and the annual energy saving ratio is over 30%. The average payback period for most cities in China is about 3.9 years. [ABSTRACT FROM AUTHOR]

The high energy consumption of data centers (DCs) has become a global issue that needs to be addressed, and as a result, their development of green rating systems (GBRSs) has been widely emphasized. Although GBRSs have been the subject of many studies for various building types, there has been few studies for DCs. This is the first study to comprehensively analyze GBRSs for DCs. The study established nine categories of standardized overall indicators and seven categories of standardized energy indicators, and furthermore reformatted and aligned the GBRS items accordingly. The study first compared GBRSs for DCs and other non-residential buildings to find out main concerns for a green DC, second compared different GBRSs for DCs to identify regional differences, and finally investigated GBRSs for DCs in China where digital infrastructure is booming. Power usage effectiveness (PUE) as a key indicator of DC energy efficiency was further analyzed. The results highlighted the importance of energy related criteria to DCs and discussed how to use the assessment criteria to appropriately guide and promote green DCs in future work. It also discussed the constraint for energy efficiency in DCs and the best practices that should be encouraged in the category of Innovation for green DCs. This study fills the gap in existing studies of GBRSs for DCs and contributes to the development of green buildings and the refinement of assessment criteria in DCs. [ABSTRACT FROM AUTHOR]

[Display omitted] • The influencing factors on energy efficiency of data center are reviewed. • Energy efficiency metrics are classified according to different priorities. • Appropriate metrics should be selected according to data center function for optimization. • Metrics are analyzed by hierarchy according to the degree of sustainable development. • Challenges and gaps are highlighted and suggestions are given for future works. The rapid development of the Internet requires data center infrastructures to be extended to a large scale with a higher power capacity and a reduction in energy consumption and carbon emissions. It is of great significance for this application to reduce energy consumption, capital cost and carbon emissions while ensuring a safe and reliable data center operation. Evaluation metrics have been considered an effective method to assess the energy efficiency of data centers and have been widely used in various data centers for many years. This work reviews the latest progress of data center evaluation metrics development in recent 20 years. Energy efficiency metrics are summarized based on energy conservation, eco-design and data center security, with the advantages and disadvantages and their correlations discussed. The applicability of each metric is analyzed. The influencing factors of energy efficiency of data centers discussed include site selection, IT equipment, power supply and distribution systems, cooling systems, building structure, and temperature control. The results show that the current evaluation metrics are valuable for the assessment of energy saving of data centers. However, more comprehensive and multi-scale evaluation metrics combining energy consumption, eco-design and data center security are needed to accelerate the progress towards carbon neutrality. In this paper, the hierarchy analysis of metrics is carried out according to sustainability, and metrics in line with sustainable development should be used preferentially. [ABSTRACT FROM AUTHOR]

A combined cooling solution is proposed to improve both thermal and energy performance for data centers with high heat density. Multi-stage heat pipe is introduced to make the internally cooled rack, which helps to illuminate the undesired mixing of hot and cold air, and makes a uniform distribution of indoor temperature. A water loop of multi cold sources is designed to make full use of waterside free cooling potentials. With the switchable and flexible operating mode, an energy efficient cooling can be expected. An operating data center in Beijing is studied and retrofitted using this solution. A comparative measurement is performed to validate the effectiveness of this combined cooling solution, which shows an improved indoor thermal environment and reduces annual cooling cost by approximate 46%. [ABSTRACT FROM AUTHOR]

• We summarized the data center energy efficiency evaluation methods and indicators, and analyzed the advantages and disadvantages of the three evaluation methods and common indicators. • We gave many suggestions to improve the current energy efficiency evaluation methods and indicators, and discussed the future directions of energy efficiency evaluation. • We analyzed the current state of energy efficiency assessment technology in the industry and academia, and made recommendations to the data center industry, decision-makers, and researchers. [Display omitted] The energy consumption by data centers is expanding in tandem with the rapid rise of the digital economy. Data centers, as high-energy-consumption organizations, have garnered extensive attention from society in order to accomplish energy conservation and emission reduction. As a result, improving the energy efficiency of cloud data centers has become a major topic of research. Researchers are working hard to develop practical energy efficiency evaluation methodologies and metrics in order to attain this goal. This article summarizes data center energy efficiency evaluation methods, classifies existing energy efficiency evaluation metrics, examines the current state and challenges of data center energy efficiency evaluation, and makes recommendations for improving energy efficiency evaluation technology to assist cloud operators, decision-makers, and researchers in developing appropriate energy efficiency evaluation strategies. We give data center researchers a better grasp of energy efficiency evaluation and encourage them to combine theory and practice in energy efficiency evaluation and utilize more advanced metrics to assess data center energy efficiency. This is a critical step in the quest for the most advanced green technology, as well as a significant step toward reaching sustainable development goals. [ABSTRACT FROM AUTHOR]

A phase change material (PCM) based cooling storage unit was designed for the emergency cooling in an air-cooled modular data center to alleviate the need for UPS while improving the operation reliability. The cooling storage unit was integrated with the ceiling air conditioner in a modular data center with the advantages of simple structure and convenient installation. Experiments including both charging and discharging process of the cooling storage unit were carried out in the constant enthalpy environment. The experimental results showed that for a 42 U standard cabinet with the size of 800 × 600 × 2000 mm (Depth × Width × Height) and the power of 6 kW, the cooling storage unit could meet the emergency cooling demand for at least 300 s at the inlet wind speed of 1.5 m/s. To further extend the cooling period, the arcuate fins were installed on the surface of the PCM plates to improve the heat transfer performance. Compared with the smooth PCM plate, the unit with arcuate fins could extend the emergency cooling time by 342 s and decrease the charging time by 1850 s under the same operating condition. [ABSTRACT FROM AUTHOR]

The growing cloud computing infrastructures increase the demand for energy-intensive machines, such as high-performance servers and the associated cooling system. That brings challenges to the energy management of the concerned modern buildings such as the data centers. Transient Thermal System Modeling (TTSM) is an essential solution to illustrate the instant thermal variation inside the buildings based on heat transformation prediction. TTSM can help plan, organize and optimize corresponding settings to improve energy efficiency. The adoption of row-based and rack-based structures as a cooling solution has recently emerged, especially boosted by its high efficiency. In contrast, few studies have concerned the related TTSM. In this study, we propose two frameworks exclusively targeting the TTSM of data centers that adopted row-based or rack-based cooling systems. A model-driven framework (MDF) and a data-driven framework (DDF) are developed, respectively. The MDF is expressed by Ordinary Differential Equations (ODEs), it is built based on the lumped-capacitance method under several pre-declared assumptions as a general TTSM solution to related structures. The DDF is developed as an extension of the Long-Short-Term-Memory (LSTM) method to realize multivariate I/O (MIMO) predictions, which is commonly demanded in Industrial Internet of Things (IIoT) applications. Both frameworks are evaluated experimentally on a physical cluster equipped with a high-level IIoT sensor measurement system. The guide for parameter-tuning of the model is also provided in the experiment part for use case analysing. The validation results show that both frameworks can build reliable TTSM in terms of real-time temperature predictions. Moreover, a comprehensive comparison study is conducted with valuable application suggestions based on experimental discoveries. [ABSTRACT FROM AUTHOR]

Highlights: [•] The energy consumption of air conditioner in a data center is modeled. [•] The model is validated by actual measured energy data. [•] Energy saving of air conditioner by increasing temperature set point is estimated. [ABSTRACT FROM AUTHOR]

Loop thermosyphon system is an efficient cooling system for energy saving in data center. It is necessary to investigate the performance of the loop thermosyphon system filled with alternative refrigerants to prompt its application. In this paper, a R32 loop thermosyphon system is experimentally investigated in an enthalpy difference laboratory for its performance. R22 and R134a are also used for comparison. The results show that the optimal filling ratio of R22, R134a and R32 is the same (37.3%). At the optimal filling ratio, the flow stability of R32 is obviously better than those of R22 and R134a. When cold water temperatures are 10–22 °C, the cooling capacity of R32 is 5.4–15.6% higher than those of R22 and R134a. After controlling extractor fans, the energy efficiency ratio of R32 is 4.9–206.6% higher than those of R22 and R134a. Moreover, the free cooling potentials of the three refrigerants in five typical climate cities across China are analyzed and compared. Compared with R22 and R134a, R32 has 3.6–11.0% and 18.5–26.7% higher free cooling percentages and annual energy efficiency ratio, respectively; and its annual average partial power usage effectiveness is 0.0012–0.0017 lower than that of other two refrigerants. [ABSTRACT FROM AUTHOR]

[Display omitted] • New hot aisle containment using a vortex flow. • Reduce the supply heat index (SHI) to 0.11 compared with the conventional layout with no containment. • 4.3% lower normalized server temperature than full hot aisle containment by relieving high backpressure. Poor airflow management attributed to hot air recirculation and cold air by-pass adds massive thermal inefficiencies to air-cooled data center's cooling system. The containment system usually was adopted to mitigate such drawbacks by isolating the space using a physical partition wall, such as hot aisle containment. Although the recirculation and by-pass can be effectively suppressed, the consequent high backpressure built up in the containment reducing the server air flowrate and possibly increasing the server temperature beyond the acceptable limit. This study presents a new enclosure mechanism for air-cooled data center using a vortex flow as an alternative to the conventional hot aisle containment to reduce the hot recirculation and, at the same time to retain the server air flowrate. The experiments conducted in a half-scale test cell (2-times smaller than a typical data center hall in length scale) examined vortex containment's effectiveness. The experiment results revealed that the average Supply Heat Index (SHI) in vortex hot aisle containment could be reduced to 0.11. Although vortex hot aisle containment did not enclose all hot recirculation compared to the full containment deployed in the benchmark layout, the relieved backpressure (the high pressure in the hot aisle) retains the rack flowrate by 42% (2.1 m/s in the vortex hot aisle containment with a diameter 1.2 m (VFL1200) layout compared to 1.2 m/s in full containment) and results in a 4.3% lower normalized server temperature (1.03 in VFL1200 layout compared to 1.076 in full containment CL). In the end, the validated CFD model provides more insights on vortex containment characteristics, such as the vortex strength and pressure distribution. [ABSTRACT FROM AUTHOR]

• A novel concept of supply and return approach temperatures is introduced • The method captures non-uniform airflow and temperature distribution in data centers. • The approach temperatures are presimulated by CFD and input into EnergyPlus. • Prototype data center energy models are developed using OpenStudio and EnergyPlus. • A case study was conducted to demonstrate the use of the data center models. Data centers in the United States consume about two percent of the nation's electricity. Because heat gains from IT equipment drive cooling demand, data centers offer unique opportunities for energy savings. However, no prototype energy model for data centers is available in the suite of existing U.S. Department of Energy's Commercial Prototype Building Models. This study presented the development of two new data center prototype models and their implementation in OpenStudio and EnergyPlus. The small-size data center model represents a computer room in a building served by computer room air conditioners (CRACs); while the large-sized model represents stand-alone data centers served by computer room air handlers (CRAHs) with a central chiller plant. For each data center model, two levels of IT equipment (ITE) load density were considered, to cover the wide range of IT power density of data centers: 40 and 100 W/ft2 (430 and 1076 W/m2) for the computer room, and 100 and 500 W/ft2 (1076 and 5382 W/m2) for the stand-alone data center. All other assumptions, such as building envelope, lighting, HVAC efficiencies and schedules, were based on the minimal requirements of ASHRAE Standard 90.1 at various vintages. We introduced a novel concept of supply and return air approach temperatures to capture the essential effects of non-uniform airflow and temperature distribution in data centers. The approach temperatures were pre-computed by computational fluid dynamics (CFD) simulations for various configurations of ITE loads and airflow containment management in data centers. A new feature was developed in EnergyPlus to implement the approach temperature method. A case study was conducted to demonstrate the use of the data center models. The two data center models cover all U.S. climate zones and can be used to evaluate energy saving measures for data centers, as well as to support development of data center energy efficiency codes and standards. [ABSTRACT FROM AUTHOR]

• An open source fast fluid dynamics (FFD) model for data center thermal management. • Validation with two simple cases and a real middle-size data center case. • Saving 53.4–58.8% of annual energy through optimal design and operation. • Reducing the computing time from 464.8 h with CFD to 7.6 h by FFD. Although computational fluid dynamics (CFD) has been widely adopted to improve data center thermal management, the high computational demand limits its applications, such as multivariate optimal design and operation. Fast fluid dynamics (FFD), which has been applied for fast airflow simulation, shows great potential. However, few research applied FFD for optimal design and operation of data center thermal management. This research improves the FFD model for data centers and conducts a comprehensive evaluation and demonstration. First, the FFD model is improved by solving the advection and diffusion equations together using an upwind scheme instead of a semi-Lagrangian advection solver in the conventional FFD model. Second, new features for data centers are added, such as a pressure correction method to simulate plenum airflow and dynamic boundary conditions for IT racks. The new FFD model is first validated with two indoor environment cases and the results show that the new FFD model has slightly better overall prediction accuracy and faster speed compared to the conventional FFD model. It is also observed that both FFD models achieve acceptable accuracy, except for a few localized disparities with experimental data, which might be due to simplified handling of turbulence viscosity near the boundaries. Furthermore, validation with a real data center shows that the FFD model achieves a similar level of accuracy as CFD when compared to the experimental measurements with some level of uncertainties. It is then demonstrated for data center optimal design and operation, which saves 53.4–58.8% of annual energy while still meeting the thermal requirements. With a much faster speed and comparable accuracy compared to CFD, the FFD model parallelized on a graphics processing unit is promising for practical model-based data center early design and operation. [ABSTRACT FROM AUTHOR]

Free cooling is commonly used in data centers to reduce cooling system energy consumption using a water-side economizer (WSE). In this system, cooling mode is switched when wet bulb temperature (T w b ) reaches switchover temperature (T s w ), and the cooling tower fans are controlled to maintain a specified cooling water approach temperature (T a p p ). Typically, both T s w and T a p p values are fixed. However, under part load conditions, T s w and T a p p can be optimized to maximize the system performance and save energy. Currently, there are few appropriate methods available to optimize T s w and T a p p in partial-free cooling mode for data centers with WSE. Thus, this study presents a model based methodology to optimize T s w and T a p p in different cooling modes. The proposed methodology is verified through a case study based on data of a real operating data center. Comparison between the calculated optimal T s w and T a p p with that derived by simulation demonstrates the satisfactory accuracies of the proposed method. The case study illustrates that when T s w and T a p p are optimized, the cooling system energy consumption can be saved by 10% when cooling load ratio is 0.6. [ABSTRACT FROM AUTHOR]

• The necessity and uniqueness of TES for data centers are discussed. • TES units and TES integrated cooling systems in data center are summarized and classified. • The designs and performance characteristics of TES units and TES integrated cooling systems in data center are analyzed. • Main trends in TES developments and applications are summarized when utilized in data center. Data center consumes a great amount of energy and accounts for an increasing proportion of global energy demand. Low efficiency of cooling systems leads to a cooling cost at about 40% of the total energy consumption of a data center. Due to specific operation conditions, high security and high cooling load is required in data center. To achieve energy saving, cost saving and high security, novel cooling systems integrated with thermal energy storage (TES) technologies have been proposed. This paper presents an extensive overview of the research advances and the applications of TES technologies in data centers. Operating conditions, energy mismatch and requirement of high security in data center were overviewed. Principles and characteristics of TES technologies were discussed. Applications of passive TES coupled air flow and applications of active TES integrated cooling system are summarizes, and the design and performance of these TES integrated thermal systems are analyzed, with a focus on energy saving, cost savings and high security. [ABSTRACT FROM AUTHOR]

This paper evaluates the potential of water-side economizers in the refrigeration system of data centers under different climate conditions. Due to the wide range of conditions along the country (Desert, Mediterranean, Temperate rainy and Tundra climate), Chile is selected as case study. The number of hours per year in which economization is possible is estimated using the data base of 22 weather stations along the country. The refrigeration system is modeled in steady state through a set of thermodynamic equations simultaneously solved using the Engineering Equation Solver (EES). The system performance is evaluated by calculating the Coefficient of Performance (COP) and the Water to Energy Ratio (WER). The latter is a new metric proposed to compare the volume of water required by the system to save 1 MWh of cooling energy. The thermodynamic analysis shows that the chiller decreases its energy usage if water-side economizers are implemented in favorable climates such as cool-summer Mediterranean with winter rain (Csc), temperate rainy (Cfb) and tundra (ET) climates. Here, the use of economizers allows a monthly increase in the COP of 50 to 120% (compared to the conventional operation) and an annual average COP ranging from 7.8 to 9.7. These climates also offer an additional gain of lower water requirements, with annual WER values ranging from 11 to 17 m3/MWh. Desert climates, on the other hand, prevent implementing economizers, offering the lowest annual average COP values (5.6–5.7). In climates in which the complete economization is impossible, the partial use of economizers allows a monthly increase in the COP of 10 to 45%. The costal influence decreases the system performance, reducing the COP and increasing the WER. [ABSTRACT FROM AUTHOR]