Data Center Containment Cooling With Predictive Airflow Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Data center containment systems face airflow mismatches due to varying IT equipment utilization, increased cooling set points, virtualization, and maintenance, leading to inefficiencies and potential overheating, as classical monitoring methods fail to accurately represent IT equipment reliability and heating rates.
Innovation Solution
The implementation of predictive control systems using passive and active flow curve methods to analyze airflow behavior, modulate cooling units, and introduce controlled leakage to maintain optimal airflow and temperature conditions, preventing CPU throttling and ensuring reliable operation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Temperature
If containment structures are sealed more perfectly to mitigate hot spots, then temperature control improves, but airflow mismatch between cooling units and IT equipment worsens
Solution Approach 1:
The system dynamically adjusts containment structure leakage characteristics based on real-time airflow measurements and predictions. Flow curves are continuously updated to reflect changing IT equipment heating rates, allowing the containment structure to adapt its airflow properties rather than maintaining a fixed sealed state. This resolves the contradiction by making the containment structure responsive to actual thermal conditions.
Solution Approach 2:
The system implements feedback control by continuously monitoring airflow through IT equipment using flow curves and comparing actual airflow against predicted requirements. When mismatches are detected, the system adjusts cooling unit operation and containment structure characteristics to restore proper airflow balance. This feedback mechanism ensures temperature control is maintained while preventing airflow mismatch.
2Loss of energy
If cooling units are tuned down to increase PUE, then energy efficiency improves, but IT equipment reliability worsens due to insufficient cooling
Solution Approach 1:
The system performs preliminary action by predicting future airflow requirements of IT equipment using flow curves before actual cooling demands arise. By anticipating heating rate changes based on equipment utilization patterns, the system pre-adjusts cooling unit operation and containment characteristics, allowing efficient operation while maintaining reliability through proactive rather than reactive control.
Solution Approach 2:
The system changes operating parameters dynamically by adjusting cooling unit setpoints, fan speeds, and containment structure characteristics based on real-time flow curve data. Rather than maintaining fixed conservative settings, the system optimizes parameters continuously to achieve the minimum cooling required for reliability while maximizing energy efficiency, resolving the trade-off between PUE and equipment operation.
3Device complexity
If external temperature sensors are used for monitoring, then system complexity is reduced, but measurement accuracy worsens because sensors cannot detect internal heating rates
Solution Approach 1:
The system introduces flow curves as an intermediary mathematical model that translates easily measurable parameters (airflow rate, pressure differential) into accurate predictions of IT equipment heating rates. Rather than directly measuring internal temperatures or heating rates with complex sensors, the flow curve intermediary enables precise indirect measurement using simple external sensors, resolving the contradiction between system simplicity and measurement accuracy.
Data Source
AI summary
A method of controlling a data center having a cold air cooling system, and at least one containment structure, comprising: determining a minimum performance constraint; determining optimum states of the cold air cooling system, a controlled leakage of air across the containment structure between a hot region and a cold air region, and information technology equipment for performing tasks to meet the minimum performance constraint, to minimize operating cost; and generating control signals to the cold air cooling system, a controlled leakage device, and the information technology equipment in accordance with the determined optimum states.