| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 3.16 MB | Adobe PDF |
Authors
Abstract(s)
Com a ascensão do mercado de Data Centers (DC), os gastos mundiais com
infraestrutura de DC devem chegar a 200 biliões de dólares em 2021, que é um aumento
de 6% em relação ao esperado para 2020 [1]. A eficiência energética nestas instalações
de elevado consumo de energia é fundamental. Já existem metodologias para mensurar
essa eficiência, como seja o índice PUE (Power Usage Effectiveness). Porém, este índice,
apesar de importante, não reflete de forma integral a eficiência. Nesta tese, são
apresentados os valores de PUE para 38 Data Centers a serem localizados em 37 cidades
brasileiras e 1 cidade (Chicago) dos Estados Unidos da América, e comparada a eficiência
deste índice com um novo criado, o índice de Eficiência no Consumo de Energia em fase
de projeto (EUED - Energy Usage Effectiveness Design). A métrica do EUED, ao invés
de considerar a potência como elemento de comparação, usa a energia. O PUE é apenas
mensurado com mais precisão no campo em forma de energia, ou seja, kWh/kWh, de
acordo com as condições externas das 8760 horas anuais. Este método possibilita ao
Investidor, com os dados do ASHRAE Weather Data Viewer, fazer simulações em
múltiplas cidades com o intuito de selecionar o local ideal para localizar um Data Center.
Este poderá ainda simular o consumo de energia anual do Data Center, considerando as
variações ao longo do ano da temperatura de bolbo seco e temperatura de bolbo húmido
coincidentes do ar. Os consumos de energia com a infraestrutura, que consistem no
somatório do consumo de energia com os equipamentos de ar condicionado,
equipamentos de Tecnologias de Informação, iluminação e outros, nas diferentes cidades
simuladas, fornece uma diferença de mais de 1,21% de São Paulo em relação a Curitiba e
de 10,61% de Rio de Janeiro em relação a Curitiba. Os índices obtidos com aplicação do
índice EUED foram respetivamente, de 1,25 [(kWh/ano)/(kWh/ano)] para Curitiba,
1,26 [(kWh/ano)/(kWh/ano)] para São Paulo e de 1,38 [(kWh/ano)/(kWh/ano)] para
Rio de Janeiro, providenciando uma diferença de 16,86% para Curitiba, de 16,19% para
São Paulo e de 10,31% para Rio de Janeiro em relação ao PUE COA (Power Usage
Effectiveness - Constant Outdoor Air). Utilizando um método transparente ao projetar
um sistema de energia com baixo impacte ambiental, o TEWI (Total Equivalent
Warming Impact), corresponde a uma métrica do impacte do aquecimento global dos
equipamentos com base no total de emissões diretas e indiretas, relacionadas a uma
correlação equivalente dos gases com efeito estufa durante a operação do equipamento e
a eliminação dos fluidos refrigerantes no fim da vida útil. Foi constatado que o consumo
anual de energia de sistemas de Aquecimento, Ventilação e Ar Condicionado é
aproximado para as duas cidades com condições climáticas similares, mas o TEWI é extremamente diferente, sendo em Chicago (EUA) de 1.781.566,043 [kg CO2/10 anos] e
em Curitiba (Brasil) de 667.042,128 [kg CO2/10 anos]. Corresponde um gasto anual em
Chicago relativamente a Curitiba de mais de 267%. Assim como o TEWI, numa visão
holística dos processos, também foi criado o índice de Impacte Total da Água (TWI -
Total Water Impact), em que os usos direto e indireto de água são utilizados para
comparar sistemas de condensação a ar e a água num sistema de refrigeração. Para tal
foram simulados chillers de condensação a ar e a água e foi obtido para uma capacidade
de 200 TR (Toneladas de refrigeração), um consumo superior de 5% no sistema com
condensação a ar e na capacidade de 500 TR, um consumo superior de 25% no sistema
de condensação a água. Ou seja, à medida que a potência de refrigeração aumenta,
verifica-se uma alteração da melhor eficiência no consumo de sistemas com chillers com
condensação a água para sistemas com chillers com condensação a ar. Porém, à medida
que surgem novos padrões, a eficiência do chiller está a tornar-se cada vez mais
importante, pois chillers de rolamentos magnéticos arrefecidos a água em ambientes
com temperatura elevado apresentam baixo valor de NPLV (Non-standard Part Load
Value), certamente levando no futuro a índices de TWI mais baixos do que os de chillers
arrefecidos a ar. De forma a simplificar a visão de eficiência num Data Center, foi criado
um novo índice denominado Data Center de Design Perfeito (PDD - Perfect Design Data
Center). Este índice facilita a visualização de um Data Center perfeito, utilizando uma
classificação em ordem crescente. Para tal, foram criadas tabelas qualitativas para
classificação de eficiência energética de DC a partir desse índice. Nas 37 cidades
simuladas, aquela que apresentou melhor índice PDD, com o valor de
0,803 [(kWh/ano)/(kWh/ano)], foi a cidade de Curitiba e aquela que apresentou o pior
índice PDD, com o valor de 0,694 [(kWh/ano)/(kWh/ano)], foi a cidade de Natal. Apesar
das simulações terem sido realizadas para as principais 37 cidades brasileiras, este
método pode ser replicado para qualquer local do globo, podendo gerar ao Investidor de
Data Center, tabelas de decisão com as simulações energéticas. Este método apresentase
como sendo o único que faz uso da psicrometria para decisão de instalação de um Data
Center, fazendo uso de sistemas free cooling, adiabáticos ou de ar condicionado com
variação do COP de acordo com a admissão do ar no condensador. O impacto dos índices
PDD e EUED para um Data Center é equivalente ao IPLV (Integrated Part Load Value),
para a escolha de um equipamento de ar condicionado.
With the rise of the Data Center market, worldwide spending on DC infrastructure is expected to reach 200 billion dollars in 2021, which is a 6% increase compared to what was expected for this year [1]. Energy efficiency in these energy-intensive installations is critical. There are already methodologies to measure this efficiency, such as the PUE (Power Usage Effectiveness) index. However, this index, although important, does not fully reflect efficiency. In this thesis, the PUE values for 38 Data Centers to be located in 37 Brazilian cities and 1 city (Chicago) in the United States of America are presented, and the efficiency of this index is compared with a new one, the Consumption Efficiency index of Energy in design phase (EUED - Energy Usage Effectiveness Design. The EUED metric, instead of considering power as a comparison element, uses energy, only after the project is completed is the PUE measured more accurately in the field in the form of energy, that is, kWh/kWh., according to the external conditions of the 8760 hours per year. This method allows the Investor, with data from ASHRAE Weather Data Viewer, to make simulations in multiple cities in order to select the ideal place to locate a Data Center. It will also be able to simulate the annual energy consumption of the Data Center, considering the variations throughout the year of the coincident dry bulb temperature and the wet bulb temperature of the air. Energy consumption with infrastructure, which consists of the sum of energy consumption with air conditioning equipment, Information Technology equipment, lighting and others, in the different cities simulated, provides a difference of more than 1.21% in São Paulo. Paulo in relation to Curitiba and 10.61% of Rio de Janeiro in relation to Curitiba. The rates obtained by applying the EUED index were, respectively, 1.25 [(kWh/year)/(kWh/year)] for Curitiba, 1.26 [(kWh/year)/(kWh/year)] for São Paulo and 1.38 [(kWh/year)/(kWh/year)] for Rio de Janeiro, providing a difference of 16.86% for Curitiba, 16.19% for São Paulo and 10.31% for Rio January in relation to the PUE COA (Power Usage Effectiveness - Constant Outdoor Air). Using a transparent method when designing an energy system with low environmental impact, the TEWI (Total Equivalent Warming Impact) corresponds to a metric of the global warming impact of equipment based on the total direct and indirect emissions, related to an equivalent correlation greenhouse gases during equipment operation and disposal of refrigerants at end of life. It was found that the annual energy consumption of Heating, Ventilation and Air Conditioning systems is approximate for the two cities with similar climatic conditions, but the TEWI is extremely different, being in Chicago (USA) 1,781,566,043 [kg CO2/10 years] and in Curitiba (Brazil) of 667,042.128 [kg CO2/10 years]. Corresponds to an annual expense in Chicago relative to Curitiba of more than 267%. As well as the TEWI, in a holistic view of the processes, the Total Water Impact index (TWI - Total Water Impact) was also created, in which the direct and indirect uses of water are used to compare air and water condensing systems in a refrigeration system. To this end, air and water condensing chillers were simulated and it was obtained for a capacity of 200 TR (Tons of cooling), a consumption higher than 5% in the system with air condensing and a capacity of 500 TR (Tons of cooling) , a consumption higher than 25% in the water condensing system. In other words, as the cooling power increases, there is a change in the best efficiency in consumption from systems with water cooled chillers to systems with air cooled chillers. However, as new standards emerge, chiller efficiency is becoming increasingly important as water-cooled magnetic bearing chillers in high temperature environments have a low NPLV (Nonstandard Part Load Value) value. certainly leading to lower TWI rates in the future than air-cooled chillers. In order to simplify the vision of efficiency in a Data Center, a new index called Perfect Design Data Center (PDD - Perfect Design Data Center) was created. This index facilitates the visualization of a perfect Data Center, using a classification in ascending order, were created qualitative tables for classification of energy efficiency of DC from this index. In the 37 cities simulated, the one that presented the best PDD index, with the value of 0.803 [(kWh/year)/(kWh/year)], was the city of Curitiba and the one that presented the worst PDD index, with the value of 0.694 [(kWh/year)/(kWh/year)], was the city of Natal. Although the simulations have been carried out for the main 37 Brazilian cities, this method can be replicated anywhere in the world, being able to generate decision tables with the energy simulations for the Data Center Investor. This method is the only one that makes use of psychrometrics to decide to install a Data Center, making use of free cooling, adiabatic or air conditioning systems with COP variation according to the intake of air in the condenser. The impact of the PDD and EUED indexes for a Data Center is equivalent to the IPLV (Integrated Part Load Value), for the choice of air conditioning equipment.
With the rise of the Data Center market, worldwide spending on DC infrastructure is expected to reach 200 billion dollars in 2021, which is a 6% increase compared to what was expected for this year [1]. Energy efficiency in these energy-intensive installations is critical. There are already methodologies to measure this efficiency, such as the PUE (Power Usage Effectiveness) index. However, this index, although important, does not fully reflect efficiency. In this thesis, the PUE values for 38 Data Centers to be located in 37 Brazilian cities and 1 city (Chicago) in the United States of America are presented, and the efficiency of this index is compared with a new one, the Consumption Efficiency index of Energy in design phase (EUED - Energy Usage Effectiveness Design. The EUED metric, instead of considering power as a comparison element, uses energy, only after the project is completed is the PUE measured more accurately in the field in the form of energy, that is, kWh/kWh., according to the external conditions of the 8760 hours per year. This method allows the Investor, with data from ASHRAE Weather Data Viewer, to make simulations in multiple cities in order to select the ideal place to locate a Data Center. It will also be able to simulate the annual energy consumption of the Data Center, considering the variations throughout the year of the coincident dry bulb temperature and the wet bulb temperature of the air. Energy consumption with infrastructure, which consists of the sum of energy consumption with air conditioning equipment, Information Technology equipment, lighting and others, in the different cities simulated, provides a difference of more than 1.21% in São Paulo. Paulo in relation to Curitiba and 10.61% of Rio de Janeiro in relation to Curitiba. The rates obtained by applying the EUED index were, respectively, 1.25 [(kWh/year)/(kWh/year)] for Curitiba, 1.26 [(kWh/year)/(kWh/year)] for São Paulo and 1.38 [(kWh/year)/(kWh/year)] for Rio de Janeiro, providing a difference of 16.86% for Curitiba, 16.19% for São Paulo and 10.31% for Rio January in relation to the PUE COA (Power Usage Effectiveness - Constant Outdoor Air). Using a transparent method when designing an energy system with low environmental impact, the TEWI (Total Equivalent Warming Impact) corresponds to a metric of the global warming impact of equipment based on the total direct and indirect emissions, related to an equivalent correlation greenhouse gases during equipment operation and disposal of refrigerants at end of life. It was found that the annual energy consumption of Heating, Ventilation and Air Conditioning systems is approximate for the two cities with similar climatic conditions, but the TEWI is extremely different, being in Chicago (USA) 1,781,566,043 [kg CO2/10 years] and in Curitiba (Brazil) of 667,042.128 [kg CO2/10 years]. Corresponds to an annual expense in Chicago relative to Curitiba of more than 267%. As well as the TEWI, in a holistic view of the processes, the Total Water Impact index (TWI - Total Water Impact) was also created, in which the direct and indirect uses of water are used to compare air and water condensing systems in a refrigeration system. To this end, air and water condensing chillers were simulated and it was obtained for a capacity of 200 TR (Tons of cooling), a consumption higher than 5% in the system with air condensing and a capacity of 500 TR (Tons of cooling) , a consumption higher than 25% in the water condensing system. In other words, as the cooling power increases, there is a change in the best efficiency in consumption from systems with water cooled chillers to systems with air cooled chillers. However, as new standards emerge, chiller efficiency is becoming increasingly important as water-cooled magnetic bearing chillers in high temperature environments have a low NPLV (Nonstandard Part Load Value) value. certainly leading to lower TWI rates in the future than air-cooled chillers. In order to simplify the vision of efficiency in a Data Center, a new index called Perfect Design Data Center (PDD - Perfect Design Data Center) was created. This index facilitates the visualization of a perfect Data Center, using a classification in ascending order, were created qualitative tables for classification of energy efficiency of DC from this index. In the 37 cities simulated, the one that presented the best PDD index, with the value of 0.803 [(kWh/year)/(kWh/year)], was the city of Curitiba and the one that presented the worst PDD index, with the value of 0.694 [(kWh/year)/(kWh/year)], was the city of Natal. Although the simulations have been carried out for the main 37 Brazilian cities, this method can be replicated anywhere in the world, being able to generate decision tables with the energy simulations for the Data Center Investor. This method is the only one that makes use of psychrometrics to decide to install a Data Center, making use of free cooling, adiabatic or air conditioning systems with COP variation according to the intake of air in the condenser. The impact of the PDD and EUED indexes for a Data Center is equivalent to the IPLV (Integrated Part Load Value), for the choice of air conditioning equipment.
Description
Keywords
Data Centers Refrigeração Índices de Desempenho Power Usage Effectiveness - PUE Energy Usage Effectiveness Design - EUED Total Equivalent Warming Impact - TEWI Perfect Design Data Center - PDD