Professional Documents
Culture Documents
a r t i c l e i n f o a b s t r a c t
Article history: Data centers cooling loads may reach about 40% of their electricity demands. The number of that type of
Received 9 November 2019 facility is continuously growing around worldwide and actions addressing energy saving are welcome to
Revised 14 November 2020
rationalize their operation. Free cooling appears as an effective solution to achieve energy savings, and the
Accepted 16 November 2020
present work proposes a methodology for assessing data center free cooling operation based on the heat
Available online 21 November 2020
exchanger approach point. The methodology can be applied to both airside and waterside economizer
Keywords: modes. The proposed methodology is applied to fourteen Brazilian state capitals that represent all coun-
Data center try climatic conditions. Free cooling environmental air temperature and humidity criteria are based on
Free cooling literature review and on ASHRAE thermal guidelines. Results show that cities such as Curitiba, São Paulo,
Air side economizer Porto Alegre and Brasília can potentially operate in both airside and waterside economizer modes for
Waterside economizer more than 30 0 0 h per year, even considering conservative data center thermal operating limits. Facilities
Approach point
placed on North and Northeast regions can only run under free cooling conditions if operational param-
ASHRAE TC 9.9 thermal guidelines
eters are to be relaxed to consider more flexible operating limits. It is shown that indirect economizers
are potentially more suitable to operate under ASHRAE recommended limits. The proposed methodology
is shaped for evaluating free cooling hours during system conception.
© 2020 Elsevier Ltd and IIR. All rights reserved.
Mots-clés: Centre de données; Refroidissement naturel; Économiseur câté air; Économiseur câté eau; Point d’approche; Directives thermiques TC 9.9 de l’ASHRAE
Abbreviation
Ap : Approach Point [◦ C]
DC : Data Center DBT : Dry Bulb Temperature [◦ C]
ESD : Electrical Static Discharge DPT : Dew Point Temperature [◦ C]
FC : Free Cooling PUE : Power Usage Effectiveness [adim]
HX : Heat Exchanger RH : Relative Humidity [◦ C]
IT : Information Technology W BT : Wet Bulb Temperature [◦ C]
WCT : Water Cooling Tower Greek letters
Symbols γ : Summation of approach points [◦ C]
(continued on next page)
∗
Corresponding author.
E-mail address: pss@mecanica.ufrgs.br (P.S. Schneider).
https://doi.org/10.1016/j.ijrefrig.2020.11.010
0140-7007/© 2020 Elsevier Ltd and IIR. All rights reserved.
E.A. Amado, P.S. Schneider and C.S. Bresolin International Journal of Refrigeration 122 (2021) 171–180
172
E.A. Amado, P.S. Schneider and C.S. Bresolin International Journal of Refrigeration 122 (2021) 171–180
173
E.A. Amado, P.S. Schneider and C.S. Bresolin International Journal of Refrigeration 122 (2021) 171–180
Table 1 A3 and A4 classes, while the older ones are more suited to operate
Recommended and Allowable ranges for DC operation.
within the A1 and A2 classes.
Adapted from ASHRAE (2015).
174
E.A. Amado, P.S. Schneider and C.S. Bresolin International Journal of Refrigeration 122 (2021) 171–180
Fig. 5. Free cooling zone considering the Recommended range limits on a psychrometric chart.
175
E.A. Amado, P.S. Schneider and C.S. Bresolin International Journal of Refrigeration 122 (2021) 171–180
ing zone considering the free cooling criteria applied to the Rec- transfer resistance is taken into account by the approach point,
ommended range limits for DBTin,max and DP Tin,max . which is the minimum temperature difference between hot and
The green color represents the zone where free cooling is cold fluid flows. The use of the approach point to refine the free-
achievable (Zone I), the red color means that the free cooling is cooling potential assessment is shown as follows applied to a wa-
not possible due to the high DBT level (Zone II), the red color rep- terside economizer with a cooling tower to be later expanded to
resents that free cooling is not possible due to the high DP T level other indirect modes.
(Zone III), and the orange zone is where free cooling is not practi- To calculate the W BTout,max , the approach point from the heat
cal due to both: DBT and DP T parameters (Zone IV). exchanger and cooling tower ApWCT are discounted from the max-
Free cooling zones for Allowable ranges limits A1, A2, A3 and imum internal dry-bulb temperature DBTin,max ,
A4 were similarly plotted as shown in Fig. 6.
W BTout,max = DBTin,max − ApHX − ApWCT (2)
3.3. Free cooling methodology for waterside cooling tower the approach values can be evaluated through an energy balance
in the heat exchanger, using the inlet and outlet temperature dif-
Differently from direct airside economizer mode, indirect ferences and effectiveness. The ApWCT is defined as the difference
modes use heat exchangers. Heat exchanger can be modeled as between the cooling tower outlet cold water and the outdoor wet-
a resistance device. To offset the resistance, the cold source tem- bulb temperature W BTout . The ApHX is defined in the cold side of
perature should be lower than the DC internal required tempera- a counter-current heat exchanger. Eq. (2) is an expression to find
ture, reducing the free-cooling available hours. That additional heat the maximum outdoor wet-bulb temperature which guarantees
176
E.A. Amado, P.S. Schneider and C.S. Bresolin International Journal of Refrigeration 122 (2021) 171–180
177
E.A. Amado, P.S. Schneider and C.S. Bresolin International Journal of Refrigeration 122 (2021) 171–180
Table 3
Capital cities assessed by country region. Weather data obtained from LABEE (2018).
ceió which achieved more than 40 0 0 h in the A1 limit. Over- were excluded from the assessment. Free cooling is possible for
all better results were found when considering more flexible the entire year for all the 14 cities when A3 or A4 limits were
limits. considered.
All the 14 assessed cities displayed at least 5900 free cool- This assessment was performed over an average of γ = 10 ◦ C.
ing hours with the waterside economizer mode when consid- Results are very sensitive to changes on γ , and W BT cumulative
ering the A2 range limits, and that number can reach at least frequency distributions can aid to assess the free cooling potential
8300 h (almost the entire year) if Fortaleza, São Luís and Belém of a given city considering variations of γ .
178
E.A. Amado, P.S. Schneider and C.S. Bresolin International Journal of Refrigeration 122 (2021) 171–180
Table 4
Free cooling potential in 14 Brazilian capitals.
Table 5
Free cooling hours while assessing only the dew-point temperature criterion and its percentage growth in relation to the assessment
considering the two criteria.
Porto Alegre RS 30◦ 02’S 3836 (+1.40%) 5323 7939 8735 (=) 8735 (=)
(+0.26%) (+0.15%)
◦
Florianópolis SC 27 36’S 1929 (+0.21%) 3760 (=) 7004 (=) 8470 (=) 8470 (=)
Curitiba PR 25◦ 26’S 5615 (+0.72%) 7299 (=) 8726 (=) 8760 (=) 8760 (=)
São Paulo SP 23◦ 33’S 4392 (+3.12%) 6425 8665 (=) 8760 (=) 8760 (=)
(+0.11%)
Rio de Janeiro RJ 22◦ 54’S 259 (+1.57%) 1134 (=) 5487 8555 (=) 8555 (=)
(+0.11%)
Vitória ES 20◦ 19’S 447 (=) 1417 (=) 5304 (=) 8494 (=) 8494 (=)
Brasília DF 15◦ 47’S 3329 (+9.94%) 5086 8745 (=) 8760 (=) 8760 (=)
(+0.14%)
Salvador BA 12◦ 58’S 17 (=) 261 (=) 5061 (=) 8729 (=) 8729 (=)
Maceió AL 9◦ 40’S 13 (+62.50%) 130 (=) 5795 (=) 8759 (=) 8759 (=)
Recife PE 8◦ 13’S 0 (=) 4 (=) 2521 (=) 8694 (=) 8694 (=)
Natal RN 5◦ 48’S 0 (=) 6 (=) 3048 (=) 8697 (=) 8697 (=)
Fortaleza CE 3◦ 43’S 0 (=) 1 (=) 221 (=) 6657 (=) 6657 (=)
São Luís MA 2◦ 32’S 0 (=) 1 (=) 168 (=) 8141 (=) 8141 (=)
Belém PA 1◦ 27’S 0 (=) 2 (=) 387 (=) 7888 (=) 7888 (=)
Table 6
Free cooling potential hours in 14 Brazilian capitals considering the average scenario (γ = 10 ◦ C).
179
E.A. Amado, P.S. Schneider and C.S. Bresolin International Journal of Refrigeration 122 (2021) 171–180
showed that Curitiba displayed the highest free cooling potential Bi, Y., Wang, Y., Ma, X., Zhao, X., 2017. Investigation on the energy saving potential
in this category, achieving 5575 economizer hours in a year, even of using a novel dew point cooling system in data centres. Energies 10, 1732.
doi:10.3390/en10111732.
considering the most conservative operating limit. Cities such as Clidaras, J., Stiver, D.W., Hamburgen, W., 2009. Water-Based Data Center.
Porto Alegre, São Paulo and Brasília also showed the potential to Patent US7525207B2; 2009. https://patentimages.storage.googleapis.com/b6/d0/
free cooling, displaying more than 30 0 0 economizer hours in those d6/d1f4357032fea5/US7525207.pdf.
Deymi-Dashtebayaz, M., Namanlo, S.V., 2019. Potentiometric and economic analysis
limits. Cities from North and Northeast did not show the potential of using air and water-side economizers for data center cooling based on vari-
to free cooling considering conservative operating limits. However, ous weather conditions. Int. J. Refrig. 99, 213–225. doi:10.1016/j.ijrefrig.2019.01.
when considering more flexible ranges such as Allowable A3 and 011.
Dino. https://www.terra.com.br/noticias/dino/com- receita- de- us- 28- bilhoes- na-
A4 limits more than 70 0 0 economizer hours were achieved. A fur-
america- latina- mercado- de- data- centers- demanda- investimento- brasileiro,
ther analysis detailing these results showed that the DP T parame- e86e110777ea555792ccb818a87d4421eedextcb.html.
ter was the critical criterion in this assessment. Driemeyer, R., 2016. Projeto de Melhoria de um Data Center com Ênfase em In-
fraestrutura e Melhoria Energética. Universidade do Vale do Taquari, Lajeado,
The waterside free cooling potential assessment was made con-
RS - Brazil Master’s thesis.
sidering a cooling tower economizer type. The criteria to free cool- Gao, X., Talebzadeh, A., Moradian, M., Han, Y., Swenson, D., Pommerenke, D., 2015.
ing proposed in this category considered the maximum supply Dependence of ESD charge voltage on humidity in data centers: Part III – esti-
air temperature (DBTin,max ) and the system approaches (ApCT and mation of ESD-related risk in data centers using voltage level extrapolation and
Chebyshev’s inequality. ASHRAE Trans. 121, 49–57.
ApHX ) as parameters to compare with the outdoor wet-bulb tem- Gözcü, O., Özada, B., Carfi, M.U., Erden, H.S., 2017. Worldwide energy analysis of ma-
perature (W BTout ). A graphical representation was developed to il- jor free cooling methods for data centers. In: Proceedings of the 16th IEEE Inter-
lustrate the criteria dependence with the sum of the system ap- society Conference on Thermal and Thermomechanical Phenomena in Electronic
Systems (ITherm), pp. 968–976. doi:10.1109/ITHERM.2017.7992592.
proaches (γ ) on a psychrometric chart. Ham, S.-W., Park, J.-S., Jeong, J.-W., 2015. Optimum supply air temperature ranges
The waterside free cooling criteria were evaluated and dis- of various air-side economizers in a modular data center. Appl. Therm. Eng. 77,
cussed considering an average scenario of γ = 10 ◦ C based on lit- 163–179. doi:10.1016/j.applthermaleng.2014.12.021.
Hydeman, M., Swenson, D., 2010. Humidity controls for data centers are they nec-
erature values. The results behaved similarly to the airside free essary. ASHRAE J. 52, 48–55.
cooling potential assessment. Curitiba displayed the highest free Islam, M.A., Mahmud, H., Ren, S., Wang, X., 2015. Paying to save: reducing cost
cooling potential in this category, achieving 6301 economizer of colocation data center via rewards. In: Proceedings of the 21st Interna-
tional Symposium on High Performance Computer Architecture (HPCA). IEEE,
hours in a year considering the most conservative operating lim-
Burlingame, CA, USA.
its. Porto Alegre, São Paulo and Brasília also showed potential LABEE, 2018. Climatic data files, 2018. https://labeee.ufsc.br/downloads/
for free cooling displaying more than 30 0 0 economizer hours in arquivos- climaticos/formato- try- swera- csv- bin.
Lee, K.-P., Chen, H.-L., 2013a. Analysis of energy saving potential of air-side free
that limit. North and Northeast cities of the country only pre-
cooling for data centers in worldwide climate zones. Energy Build. 64, 103–112.
sented the potential to free cooling considering Allowable A2 limits doi:10.1016/j.enbuild.2013.04.013.
onwards. Lee, K.-P., Chen, H.-L., 2013b. Analysis of energy saving potential of air-side free
cooling for data centers in worldwide climate zones. Energy Build. 64, 103–112.
doi:10.1016/j.enbuild.2013.04.013.
Declaration of Competing Interest Malkamäki, T., Ovaska, S.J., 2012. Solar energy and free cooling potential in European
data centers. Proc. Comput. Sci. 10, 10 04–10 09. doi:10.1016/j.procs.2012.06.138.
The authors declare that they have no known competing finan- Marcinichen, J.B., Olivier, J.A., Thome, J.R., 2012. On-chip two-phase cooling of dat-
acenters: cooling system and energy recovery evaluation. Appl. Therm. Eng. 41,
cial interests or personal relationships that could have appeared to 36–51. doi:10.1016/j.applthermaleng.2011.12.008.
influence the work reported in this paper. Oró, E., Depoorter, V., Garcia, A., Salom, J., 2015. Energy efficiency and renewable
energy integration in data centres. Strategies and modelling review. Renew. Sus-
tain. Energy Rev. 42, 429–445. doi:10.1016/j.rser.2014.10.035.
Acknowledgments Rong, H., Zhang, H., Xiao, S., Li, C., Hu, C., 2016. Optimizing energy consumption for
data centers. Renew. Sustain. Energy Rev. 58, 674–691. doi:10.1016/j.rser.2015.
Amado acknowledges Coordenação de Aperfeiçoamento de Pes- 12.283.
Santos, A.F., Gaspar, P.D., Souza, H.J.d., 2019. Evaluation of the heat and en-
soal de Nível Superior (CAPES) for his MSc grant; Smith Schneider
ergy performance of a datacenter using a new efficiency index: energy us-
acknowledges Conselho Nacional de Desenvolvimento Científico e age effectiveness design - eued. Braz. Arch. Biol. Techonol. 62. doi:10.1590/
Tecnológico for his research grant (CNPq-PQ 301619/2019-0). 1678- 4324- smart- 2019190021.
Statista, 2019. Worldwide it Spending Data Center Systems 2012–2020 Statis-
tic. Available at: <https://www.statista.com/statistics/496373/datacenter-
References
colocation-revenueworldwide/>, Statista, 2019.
Shende, S., 2015. Colocation Market by Type (retail Colocation and Wholesale Colo-
Amoabeng, K.O., Choi, J.M., 2016. Review on cooling system energy consumption cation), End-User (Large Enterprises and Small and Medium Enterprises) and
in internet data centers. Int. J. Air-Cond. Refrig. 24 (04), 1630 0 08. doi:10.1142/ Industry Vertical (BFSI, Government & Public Sector, Telecom & IT, healthcare
S201013251630 0 081. & Life Sciences, Energy) – Global Opportunity Analysis and Industry Forecast,
ASHRAE, 2015. ASHRAE Thermal Guidelines for Data Processing Environments, 2014–2020.
fourth American Society of Heating, Refrigerating and Air-Conditioning Engi- Siriwardana, J., Jayasekara, S., Halgamuge, S.K., 2013. Potential of air-side economiz-
neers, Inc, Atlanta, GA. ers for data center cooling: a case study for key Australian cities. Appl. Energy
ASHRAE, 2016a. ASHRAE Handbook: HAVC Systems and Equipment, SI Edition. 104, 207–219. doi:10.1016/j.apenergy.2012.10.046.
American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc, The Green Grid. Updated air-side free cooling maps: the impact of ASHRAE 2011 Al-
Atlanta, GA. lowable Ranges White Paper #46. 2012. Available at https://datacenters.lbl.gov/
ASHRAE, 2016b. Data Center Power Equipment Thermal Guidelines and Best Prac- sites/default/files/WP46UpdatedAirsideFreeCoolingMapsTheImpactofASHRAE2011
tices. Technical Report. American Society of Heating, Refrigerating and Air-Con- AllowableRanges.pdf.
ditioning Engineers, Inc. The Green Grid, 2009. Air-Side Free Cooling Maps.
Avgerinou, M., Bertoldi, P., Castellazzi, L., 2017. Trends in data centre energy con- Wan, F., Swenson, D., Hillstrom, M., Stayer, C., Pommerenke, D., 2013. The effect of
sumption under the european code of conduct for data centre energy efficiency. humidity on static electricity induced reliability issues of ICT equipment in data
Energies 10 (10), 1470. doi:10.3390/en10101470. centers – motivation and setup of the study. ASHRAE Trans. 119, 341–358.
Beaty, D.L., Quirk, D.P., Morrison, F.T., 2019. Designing data center waterside econo- Yuventi, J., Mehdizadeh, R., 2013. A critical analysis of power usage effectiveness
mizers. ASHRAE J. 57–61. and its use in communicating data center energy consumption. Energy Build.
Belizário, A.C., 2018. Avaliação Energética e Financeira para Utilização de Sistemas 64, 90–94.
de ar Condicionado Acionados por Energia Solar em Ambientes de Missão Zhang, H., Shao, S., Xu, H., Zou, H., Tian, C., 2014. Free cooling of data centers: a
Crítica Para Diferentes Regiões Climáticas. Universidade de São Paulo, São Paulo review. Renew. Sustain. Energy Rev. 35, 171–182. doi:10.1016/j.rser.2014.04.017.
Ph.D. thesis.
180