Server room CFD analyses results

Files included in this dataset contain transient records of the most important variables of results of 9 scenarios of Poznan Supercomputing and Networking Center testbed CFD analysis.

Modeled testbed included server room equipped with two heterogeneously occupied racks (with each server modeled separately) and one roof-mounted cooling device. Setup was examined virtually for three servers power usage scenarios (50%, 75% and 100% of maximum value). For major part of active servers the power usage was changed respectively in each scenario, while some of them preserved daily-averaged values (inactive devices are also included). For each power usage level also three different air conditioning temperature setups (24, 27 and 30 Celsius degrees) were tested. The most relevant information about the model can be found below:

• server room dimensions along x, y and z axis (with one of the corners placed in the origin): (4.86, 4.00, 3.00) [m]
• racks dimensions: (0.6,1.18,2.1),(0.8,1.18,2.2) [m]
• racks position (translation from 0,0,0): (1.4,1.95,0), (2.08,1.95,0) [m]
• first rack equipment (position [-], size [-], airflow [m³/s], power usage (50%/75%/100% scenario) [W], flow orientation [-]):
  • 0, plug, -, -/-/-, -,
  • 1, plug, -, -/-/-, -,
  • 2, 4U, 0.0, 0.0/0.0/0.0, y,
  • 6, 2U, 0.0, 0.0/0.0/0.0, y,
  • 9, plug, -, -/-/-, -,
  • 10, 2U, 0.05, 230/345/460, y,
  • 20, 2U, 0.025, 90/90/90, y,
  • 22, 2U, 0.05,  450/675/900, y,
  • 24, 4U, 0.12, 1000/1500/2000, y,
  • 28, 2U, 0.05,  230/345/460, y,
  • 30, 4U, 0.12, 1000/1500/2000, y,
  • 34, 2U, 0.05,  400/600/800, y,
  • 36, 2U, 0.05,  400/600/800, y,
• second rack equipment (position [-], size [-], airflow [m³/s], power usage (50%/75%/100%  scenario) [W], flow orientation [-]):
  • 0, 3U, 0.0, 0.0/0.0/0.0, y,
  • 3, 3U, 0.0, 0.0/0.0/0.0, y,
  • 6, plug, -, -/-/-, -,
  • 7, plug, -, -/-/-, -,
  • 8, 3U, 0.0, 0.0/0.0/0.0, y,
  • 11, 2U, 0.0, 0.0/0.0/0.0, y,
  • 13, 2U, 0.025, 69/69/69, x,
  • 15, 3U, 0.0/0.0/0.0, 0.0/0.0/0.0, y,
  • 18, 3U, 0.15, 1350/2025/2700, x,
  • 22, 2U, 0.05, 161/161/161, y,
  • 24, 2U, 0.025, 69/69/69, x,
  • 26, 2U, 0.0, 0.0/0.0/0.0, y,
  • 28, 2U, 0.025, 23/23/23, x,
  • 30, 2U, 0.05, 258.5/258.5/258.5, y,
  • 34, plug, -, -/-/-, -,
  • 35, plug, -, -/-/-, -,
  • 36, plug, -, -/-/-, -,
  • 37, plug, -, -/-/-, -,
  • 38, plug, -, -/-/-, -,
  • 40, plug, -, -/-/-, -,
  • 41, plug, -, -/-/-, -,
  • air conditioner max. cooling capacity: 10 [kW]
  • air conditioner airflow: 0.533 [m3/s]
  • air conditioner inlet center: (3.76, 1.3, 3.0)
  • air conditioner outlet center: (3.76, 2.8, 3.0)
  • ventilation airflow: 0.0252 [m3/s]
  • ventilation inlet center: (2.8125, 3.6475, 3.0)
  • ventilation outlet center: (2.95, 0.605, 3.0)
• domain cuboid cavities (start point, end point):
  • (0, 0, 3.0), (0.72, 4.1, 3.27),
  • (0.32, 0, 3.27), (0.72, 3.78, 3.5)
• domain cuboid obstacles (start point, end point):
  • (3.61, 0, 0), (4.86, 0.44, 1.89)
  • (0.32, 3.78, 0), (0.72, 4.1, 3.0)

The dataset contains records of below variables:

• temperature from four virtual probes placed at the same locations as real sensors in the testbed:
  •   (1.8, 2.12, 1.42) - close to inlet of a server from the first rack
  •   (4.8, 1.65, 1.5) - close to air conditioner wall-mounted control panel
  •   (2.5, 2.09, 0.95) - close to inlet of a server from the second rack
  •   (2.8, 0.65, 0.75) - probe distanced from both racks and air conditioner, imitating "neutral" point
• surface-averaged temperature at inlet of each server
• surface-averaged temperature at outlet of each server
• surface-averaged temperature at air conditioner inlet
• surface-averaged temperature at air conditioner outlet