Data Center Context Trends & Energy Use Joshua Hatch Director, Sustainability Analytics Brightworks Sustainability Advisors
Data Center Context Societal trends in business and individuals: Connectivity Ever faster bandwidth Wider availability More devices Streaming content Data and more Data Increasing Availability Increasing Quantity & Storage Increasing Reliance
Data Center Context Data centers are hidden behind: Living in the Cloud Storage Email Smart Phones Banking Calendars Leisure browsing ecommerce Shopping Entertainment Streaming Content Social News Business Health Care Government Telecommunication
Data Center Context Source:Koomey, Jonathon. Growth in Data Center Electricity Use 2005 2010. Analytics Press. 2011.
Data Center Context Worldwide, data center electrical energy use: DOUBLED from 2000 2005 56% Increase from 2005 2010 Still, in 2010, they only represent: 1.3% of global electricity use 2% of US electricity use SOURCE: Koomey, Jonathon. Growth in Data Center Electricity Use 2005 2010. Analytics Press. 2011.
Data Center Context Types Enterprise Large owner/operator data centers Co-location Landlord/tenant, BYO-servers Cloud Data as service, rent bandwidth In-house Server closets in offices Key Considerations Uptime is king: 24 / 7 / 365 Redundancy Backup power Cooling-driven
Data Center Context Where do data centers want to be located? Cheap electricity High grid reliability Cheap large land plots Strong network/internet connection What do they look like? Large tilt up warehouses or renovations of industrial buildings Lots electrical lines, often cooling towers Rooftops packed with mechanical systems
Data Center Context Prioritizing Energy Efficiency Opportunities Location & Climate Electrical Distribution Servers Cooling Airflow Optimization Relevance in Oregon Inexpensive, high quality electricity No sales tax, economic development incentives Advantageous climate for free cooling
PART I: Overall Process, Model Geometry, and Internal Loads Aryn Bergman, P.E. TL CIRCLE
What s Important Differences b/t LEED and the ETO processes Establish the baseline Servers, cooling equipment, power distribution Support areas Depends on datacenter loads TL CIRCLE
Overall Workflow TL CIRCLE
Model Geometry TL CIRCLE
Calculating Power Distribution Efficiencies TL CIRCLE
Calculating Server Loads TL CIRCLE
Inputting Datacenter Loads into equest TL CIRCLE
Example EEM Savings TL CIRCLE
PART II: MODELING THE HVAC SYSTEMS BRIAN GOLDSMITH, LEED AP ALFA TECH
Data Center LEED Baseline Model Established by ASHRAE 90.1 90.1-2007 not suited for data centers Section G3.1.1 Exception equest is good for baseline model Usually just DX packaged units w/ little complexity Requires few or no spreadsheets
Data Center ETO Baseline Model Established by ETO New Buildings Program Technical Guidelines Dictated by State Energy Efficiency Design (SEED) Appendix L Based on ASHRAE 90.1-2007 Baseline model depends on proposed design Has provisions for evaporative cooling http://energytrust.org/library/forms/nbe_pg_techguidelines.pdf
Data Center Proposed Model equest cannot model modern data centers Limitations: SAT>70 High economizer cutoff Complicated Controls Evap cooling This is where Excel spreadsheets come in
Design Model Spreadsheets Excel is more powerful for these applications Add-in: PsychATCE VBA scripting and macros PsychATCE =RHfromDBWB(D4,E4,Elev) =DPfromDBWB(D4,E4,Elev) 8,760 hour TMY weather data from equest website Doe2.com/index_wth.html
Psych Chart Weather Data and Operating Modes 50 yr max wetbulb A:Mix OA/RA w/ Humid. B:100% OA w/ EC/H C:100% OA D:100% OA w/ Evap. Cool E1:100% OA w/ Evap. Cool and high DP E2:100% OA w/ Mech. Cool F: Mix OA/RA w/ high DP G:Mix OA/RA Data center operating criteria
Another Data Center Example Operating Window Max Wetbulb
Spreadsheet Input Screen Macros and buttons Almost all these values are calculated using PsychATCE Taken from schedules at full load not calculated Effectiveness used for evaporative cooling calc
Spreadsheet Calculation and Controls Sequence Supply Temp Reset Equation =IF(Y4,LADB_MIN,IF(OR(Z4,AB4,AC4),MIN(DBfromWBDP(E4,LADP_MIN,Elev)+FAN_HEAT,LADB_ MAX),IF(AA4,LADB_MAX,MAX(LADB_MIN,DBfromDPRH(H4,LARH_MAX,Elev)+FAN_HEAT))))
Detailed Calculation Example This is where evap. media effectiveness is factored in
Spreadsheet Output Screen Results broken down by operating mode No Mech Cooling Input into equest Calibrate equest to match
CLOSING POINTS Efficient servers and distribution systems have the most potential for savings Spreadsheet calculations can be used to demonstrate these savings Revit is a powerful tool for building modeling Be patient but persistent with LEED Learn how to use Excel Understanding psychrometrics and the controls sequence is key With the right tools and methodology, energy modeling can be used to improve data center mechanical system design efficiency.