Modular Coolant Distribution Units for Low-Drop Data Center Cooling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High power computing systems generate significant heat, requiring high coolant flow rates that lead to pressure drops and wasted energy in liquid cooling systems, and existing systems lack modularity and redundancy for effective thermal management.
Innovation Solution
A modular coolant distribution unit (CDU) with independent cooling loops and multiple modular CDUs (mCDUs) that allow for independent control of fluid flow, redundancy, and easy reconfiguration, along with heat exchangers to transfer heat efficiently between loops.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Temperature
If high coolant flow rates are used to cool high power computing systems, then heat removal effectiveness is improved, but pressure drops and energy waste increase
Solution Approach 1:
The cooling system is divided into multiple independent modular CDUs, each handling a portion of the total cooling load. This segmentation allows each module to operate at optimized flow rates, reducing overall pressure drops and energy consumption while maintaining effective heat removal across the entire system.
Solution Approach 2:
The system incorporates dynamically adjustable flow control mechanisms within each modular CDU, allowing real-time optimization of coolant flow rates based on actual thermal loads. This dynamic adjustment ensures efficient heat removal while minimizing energy waste from excessive flow rates.
2Temperature
If high coolant flow rates are used to cool high power computing systems, then heat removal effectiveness is improved, but pressure drops increase
Solution Approach 1:
The cooling system is divided into multiple independent modular CDUs, each handling a portion of the total cooling load. This segmentation allows each module to operate at optimized flow rates, reducing overall pressure drops and energy consumption while maintaining effective heat removal across the entire system.
3Device complexity
If traditional non-modular cooling systems are used, then system simplicity is maintained, but thermal management effectiveness and redundancy are reduced
Solution Approach 1:
The cooling system is divided into multiple independent modular CDUs, each handling a portion of the total cooling load. This segmentation allows each module to operate at optimized flow rates, reducing overall pressure drops and energy consumption while maintaining effective heat removal across the entire system.
Solution Approach 2:
The system incorporates dynamically adjustable flow control mechanisms within each modular CDU, allowing real-time optimization of coolant flow rates based on actual thermal loads. This dynamic adjustment ensures efficient heat removal while minimizing energy waste from excessive flow rates.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The system provides effective thermal management with reduced pressure drops, improved energy efficiency, and continuous operation even with mCDU malfunctions, ensuring optimal cooling of data center components.
Implementation Method 1
a heat exchanger configured to cool the rack cooling fluid
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The technology is directed to a modular liquid-cooling information technology ("IT") system, apparatus, and method for cooling and thermal management of IT equipment in a data center. An improved coolant distribution unit (CDU) provides cooling fluid to data racks in a data center and includes a plurality of modular CDUs ("mCDUs"). The mCDUs provide for modularity of the liquid-cooling system and greater control over fluid flow through the CDU and to the data racks. Additionally, methods are disclosed for controlling components within each mCDU to allow for continual modifications to fluid flow through the system to provide optimal cooling of data racks.