Data Center Containment Cooling Control for Airflow Mismatch
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Data centers face airflow mismatches and inefficiencies due to containment solutions, leading to increased operational costs and reduced reliability of IT equipment, as external temperature sensors fail to account for internal heating rates and airflow reduction.
Innovation Solution
A predictive control system using passive and active flow curves to analyze airflow behavior and temperature correlations, integrated with a central controller to modulate cooling units and containment structures, ensuring optimal airflow and temperature management for IT equipment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If containment solutions are implemented to seal the contained space, then energy efficiency is improved, but airflow mismatch between cooling units and IT equipment occurs
Solution Approach 1:
The system performs preliminary actions by predicting future airflow requirements of IT equipment based on workload patterns and historical data. The predictive model anticipates cooling demands before they occur, allowing the containment system to pre-adjust airflow paths and prevent airflow mismatches before they happen, thus maintaining both energy efficiency and reliability
Solution Approach 2:
The system implements feedback mechanisms by continuously monitoring actual airflow conditions, IT equipment temperatures, and cooling unit performance. This real-time feedback is fed into the predictive model to refine future predictions and adjust containment airflow dynamically, ensuring the system maintains optimal performance while sealed
2Loss of energy
If cooling units are tuned down to increase PUE, then energy efficiency is improved, but IT equipment reliability deteriorates due to insufficient cooling
Solution Approach 1:
The system applies dynamics by making the containment airflow system adaptive and flexible. Rather than fixed cooling unit settings, the system dynamically adjusts airflow distribution based on real-time predictions of IT equipment cooling needs. This allows cooling units to operate at optimal efficiency levels while ensuring adequate cooling is delivered when and where needed, maintaining both PUE and reliability
Solution Approach 2:
The predictive model performs preliminary assessment of cooling requirements based on IT workload patterns, allowing the system to prepare and allocate cooling resources in advance. This prevents situations where cooling units are insufficiently tuned, as the system proactively ensures adequate cooling capacity is available before thermal issues arise
3Device complexity
If external temperature sensors are used for monitoring, then measurement simplicity is improved, but measurement precision deteriorates as sensors become agnostic to internal heating rates
Solution Approach 1:
The system introduces an intermediary predictive model that acts as a bridge between external temperature sensors and the actual internal thermal conditions of IT equipment. This model uses multiple inputs including external sensor data, IT workload information, and historical thermal patterns to infer internal heating rates and temperatures, providing precise measurements without requiring direct internal sensors in every equipment component
Solution Approach 2:
The system replaces direct physical temperature sensing (mechanical approach) with a predictive computational model that calculates internal temperatures based on workload and environmental data. This substitution of mechanical sensing with intelligent prediction maintains measurement simplicity while achieving high precision by accounting for internal heating rates that external sensors cannot directly measure
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.