Adaptive CRAC Group Control for Data Center Hot Spot Response
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Data centers face inefficiencies in cooling systems due to oversized setups that consume excessive energy, with existing control logic often being non-automatic, specific to certain equipment, and failing to address hot spots effectively, leading to unsatisfactory thermal conditions.
Innovation Solution
A smart adaptive cooling system control algorithm using the Rack Cooling Index Over Temperature (RCIHI) to determine hot spots and dynamically adjust cooling requests among computer room air conditioners (CRACs), allowing for independent operation and automatic control without human intervention, optimizing HVAC unit status, return air temperature, and supply fan speed to maintain desired thermal levels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If cooling system is oversized and runs excessively to keep data center cool, then thermal safety is guaranteed, but cooling energy consumption skyrockets
Solution Approach 1:
The system dynamically adjusts CRAC unit operation based on real-time thermal conditions. Each CRAC unit independently controls its operation (on/off status, return air temperature, supply fan speed, cooling valve) based on current hot spot detection, transitioning from static oversized operation to dynamic demand-responsive operation that maintains thermal safety while reducing energy consumption.
Solution Approach 2:
The system implements feedback control by continuously monitoring thermal conditions through RCIHI calculation and adjusting CRAC operations accordingly. When hot spots are detected (RCIHI exceeds threshold), cooling requests are generated; when thermal conditions improve, requests are recalled, creating a closed-loop feedback system that optimizes energy use while maintaining thermal safety.
2Reliability
If existing control logic uses single HVAC unit to address hot spots, then some cooling is provided, but thermal conditions remain unsatisfactory
Solution Approach 1:
The system merges the cooling capabilities of multiple CRAC units to address hot spots. When a hot spot is detected, cooling requests are generated not only for the local CRAC but also propagated to other CRAC units in the network, combining their cooling outputs to more effectively eliminate hot spots and improve thermal conditions.
Solution Approach 2:
The control system is segmented into independent CRAC units, each running the control algorithm independently. This segmentation allows distributed decision-making where each unit can autonomously respond to local thermal conditions while coordinating with others through cooling requests, improving overall cooling efficiency and responsiveness.
3Ease of operation
If existing control logic requires human intervention, then manual adjustments can be made, but automation level is insufficient
Solution Approach 1:
The system implements self-service automation where each CRAC unit independently executes the control algorithm without human intervention. The units autonomously detect hot spots, generate cooling requests, adjust their operations, and recall requests when appropriate, fully automating the thermal management process while eliminating the need for manual control.
4Reliability
If existing control logic is specific to certain equipment types, then specialized control can be implemented, but adaptability to different data centers is limited
Solution Approach 1:
The control algorithm is designed as a universal solution that can be deployed across different data center configurations and equipment types. Each CRAC unit independently runs the same algorithm, which adapts to local conditions through real-time RCIHI calculation and dynamic cooling request generation, providing both equipment-specific precision and broad adaptability without requiring customization for different data center types.
Data Source
AI summary
Described is an adaptive automatic computer room air conditioner (CRAC) or computer room air handler (CRAH, CRAC and CRAH is referred interchangeably in this article) group control method. This method automatically controls each HVAC unit's on/off status, return temperature set point, fan speed or cooling valve position to secure the data center thermal environment for server's secure running and minimize the cooling energy use. The method creates a comprehensive feedback control loop between temperature sensor network, data center environment and CRACs.