Thermal Modeling of Data Centers for Control and Energy Usage Optimization

Athavale, Jayati; Yoda, Minami; Joshi, Yogendra

doi:10.1016/bs.aiht.2018.07.001

Cited by 18 publications

(4 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Traditional approaches for thermal modeling of data centers include computational fluid dynamics (CFD) models and simplified physical models. 7 CFD-based temperature models can accurately simulate and evaluate the thermal distribution in the rack room. However, this method has high computational overhead and a complex modeling process, which is unsuitable for real-time thermal management.…”

Section: Introductionmentioning

confidence: 99%

A multi-setpoint cooling control approach for air-cooled data centers using the deep Q-network algorithm

Chen,

Guo,

Liu

et al. 2024

Measurement and Control

View full text Add to dashboard Cite

Cooling systems provide a safe thermal environment for the reliable operation of IT equipment in data centers (DCs) while generating significant energy consumption. Therefore, to achieve energy savings in cooling system control under dynamic thermal distribution in DCs, this paper proposes a multi-setpoint cooling control approach based on deep reinforcement learning (DRL). Firstly, a thermal model based on the XGBoost algorithm is constructed to precisely evaluate the thermal distribution in the rack room to guide real-time cooling control. Secondly, a multi-set point cooling control approach based on the deep Q-network algorithm (DQN-MSP) is designed to finely regulate the supply air temperature of each air conditioner by capturing the thermal fluctuations to ensure the dynamic balance of cooling supply and demand. Finally, we adopt the extended CloudSimPy simulation tool and the real workload trace of the PlanetLab system to evaluate the effectiveness and performance of the proposed approach. The simulation results show that the proposed control solution effectively reduces the cooling energy consumption by over 2.4% by raising the average air supply temperature of the air conditioner while satisfying the thermal constraints.

show abstract

Section: Introductionmentioning

confidence: 99%

A multi-setpoint cooling control approach for air-cooled data centers using the deep Q-network algorithm

Chen,

Guo,

Liu

et al. 2024

Measurement and Control

View full text Add to dashboard Cite

show abstract

“…Approximately half of this energy was spent on cooling these centers. [ 7 ] Advanced cooling technology that directly cools integrated circuits (ICs) (rather than the entire data center) offers an alternative. Furthermore, IC miniaturization has increased the power density of a packaged chip to 10 6 W cm −2 in 2018; power densities continue to rise, and often induce highly nonuniform heat generation.…”

Section: Introductionmentioning

confidence: 99%

Electrocaloric Effect of Perovskite High Entropy Oxide Films

Son

Zhu

Trolier‐McKinstry

2022

Adv Elect Materials

View full text Add to dashboard Cite

Such small-scale cooling is also of interest for on-chip cooling of semiconductor circuits. [6] In 2020, data centers in the United State used an estimated 75 billion kWh, more than 2% of the nation's total energy consumption. Approximately half of this energy was spent on cooling these centers. [7] Advanced cooling technology that directly cools integrated circuits (ICs) (rather than the entire data center) offers an alternative. Furthermore, IC miniaturization has increased the power density of a packaged chip to 10 6 W cm −2 in 2018; power densities continue to rise, and often induce highly nonuniform heat generation. [8] It has been reported that more than 50% of IC failures are related to thermal control; this rate is expected to increase with further miniaturization. [9] Therefore, new approaches for local cooling of IC are needed. [10] Solid-state cooling, a potential solution, is highly compact, efficient, and environmentally friendly. Research involving thermoelectric, electrocaloric, and magnetocaloric effects is promising. [11][12][13] In 2006, large electrocaloric effects were observed in PZT thin films. [14] Unlike magnetocaloric and barocaloric materials, the electrocaloric effect does not require bulky external magnets or pressure apparatus. Electrocalorics enable relatively good energy efficiency that can be further improved via charge recovery. [15,16] Achieving large temperature changes in the electrocaloric material necessitates inducing large field-induced entropy changes. For this purpose, relaxor ferroelectrics are of special interest due to the disordered polarization and polar nanoregions. [17] A potential additional contribution to the entropy change in a ferroelectric-based material is to use high compositional complexity, e.g. through entropy engineering to further disorder the polarization on a local scale. In 2004, entropy stabilization of metallic alloys was introduced. [18,19] In 2015, the concept was broadened to include entropy stabilized oxides. [20] Various entropy stabilized oxides (ESO) have shown high dielectric constants, superior ionic conductivity, and extremely high-temperature stability; that is, entropy stabilization provides an additional approach to property tunability with respect to enthalpy-based compositional design. [21][22][23] In cases where reversible phase transformations have not yet been demonstrated, potential ESO may be called high entropy oxides (HEO). [24] Triggered by the novel concept, several attempts to investigate the electrocaloric effect (ECE) based on HEO have been made. A-site disordered HEO ceramics show good This paper describes two perovskite high entropy oxide (PHEO) compositions: Pb(Hf 0.2 Zr 0.2 Ti 0.2 Nb 0.2 Mn 0.2 )O 3 (Mn PHEO) and Pb(Hf 0.2 Zr 0.2 Ti 0.2 Nb 0.2 Al 0.2 )O 3 (Al PHEO). Powders are prepared by conventional solid state sintering by first pre-reacting the B-site oxides, then adding PbO. Phase pure Mn PHEO powder is obtained following calcination of the mixed powders at 750 °C for 240 min; however, secondary phases persist...

show abstract

“…Benchmarking studies (Stansberry and Kudritzki, 2012) reveal that cooling infrastructure consumes 30%−50% of the total power in a data center; in the worst-case scenarios, the facility-side power consumption exceeds that for the IT equipment. These alarming trends, when coupled with increasing energy costs and associated environmental impact, motivate much of the current research in this area, which emphasizes the need for energy-efficient thermal management schemes, without compromising server reliability (Athavale et al , 2018b).…”

Section: Introductionmentioning

confidence: 99%

Genetic algorithm based cooling energy optimization of data centers

Athavale

Yoda

Joshi

2021

HFF

Self Cite

View full text Add to dashboard Cite

Purpose This study aims to present development of genetic algorithm (GA)-based framework aimed at minimizing data center cooling energy consumption by optimizing the cooling set-points while ensuring that thermal management criteria are satisfied. Design/methodology/approach Three key components of the developed framework include an artificial neural network-based model for rapid temperature prediction (Athavale et al., 2018a, 2019), a thermodynamic model for cooling energy estimation and GA-based optimization process. The static optimization framework informs the IT load distribution and cooling set-points in the data center room to simultaneously minimize cooling power consumption while maximizing IT load. The dynamic framework aims to minimize cooling power consumption in the data center during operation by determining most energy-efficient set-points for the cooling infrastructure while preventing temperature overshoots. Findings Results from static optimization framework indicate that among the three levels (room, rack and row) of IT load distribution granularity, Rack-level distribution consumes the least cooling power. A test case of 7.5 h implementing dynamic optimization demonstrated a reduction in cooling energy consumption between 21%–50% depending on current operation of data center. Research limitations/implications The temperature prediction model used being data-driven, is specific to the lab configuration considered in this study and cannot be directly applied to other scenarios. However, the overall framework can be generalized. Practical implications The developed framework can be implemented in data centers to optimize operation of cooling infrastructure and reduce energy consumption. Originality/value This paper presents a holistic framework for improving energy efficiency of data centers which is of critical value given the high (and increasing) energy consumption by these facilities.

show abstract

Thermal Modeling of Data Centers for Control and Energy Usage Optimization

Cited by 18 publications

References 31 publications

A multi-setpoint cooling control approach for air-cooled data centers using the deep Q-network algorithm

A multi-setpoint cooling control approach for air-cooled data centers using the deep Q-network algorithm

Electrocaloric Effect of Perovskite High Entropy Oxide Films

Genetic algorithm based cooling energy optimization of data centers

Contact Info

Product

Resources

About