Data Center Cooling Redundancy Computation Method

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional methods for computing cooling redundancy in data centers are computationally complex and inefficient, often requiring high computational complexity and resulting in over-provisioning or inadequate cooling, especially when calculating real-time cooling redundancy across multiple racks and coolers.

Innovation Solution

A computer-implemented method and apparatus for determining cooling redundancy by modeling data center layouts, computing airflow fractions for each rack, sorting and summing cooler contributions, and reporting redundancy as a value, which can include calculating surplus cooling factors and separating airflow contributions for hot and cold aisles.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If conventional methods are used to compute cooling redundancy, then comprehensive cooling coverage is achieved, but computational complexity increases to 2^n

Engineering Contradiction:
Improvecooling redundancyVSAvoidcomputational complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent segments the cooling redundancy computation by separating cold aisle and hot aisle calculations, and further dividing into capture index computation and redundancy determination phases. This segmentation reduces the overall computational complexity from exponential 2^n to manageable polynomial complexity by breaking down the monolithic calculation into independent, smaller computational tasks.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary computations of capture indices between coolers and racks before computing actual cooling redundancy. By pre-calculating airflow capture relationships and storing them as coefficients, the system avoids re-computing these relationships during redundancy analysis, significantly reducing the computational burden when assessing cooling adequacy under various failure scenarios.

Inventive Principle:
Principle #10Preliminary action

2Productivity

If real-time cooling redundancy assessment is implemented, then cooling efficiency is improved, but computational resources are consumed

Engineering Contradiction:
Improvecooling assessment efficiencyVSAvoidcomputational resource consumption
Core Design Contradiction:
ProductivityVSUse of energy by moving object

Solution Approach 1:

The system pre-computes capture indices and stores them for rapid retrieval during real-time redundancy assessments. This preliminary calculation phase separates the computationally intensive geometric and airflow relationship analysis from the time-critical redundancy evaluation, enabling real-time assessment with minimal computational resource consumption during operation.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The computation is segmented into offline preparation phases (calculating capture indices, building coefficient matrices) and online assessment phases (evaluating redundancy given pre-computed data). This allows the system to perform comprehensive real-time redundancy assessments by combining pre-computed segments rather than recalculating everything from scratch, reducing real-time computational resource demands.

Inventive Principle:
Principle #1Segmentation

3Reliability

If cooling redundancy is over-provisioned, then reliability is improved, but cost increases

Engineering Contradiction:
Improvecooling reliabilityVSAvoidcooling capacity
Core Design Contradiction:
ReliabilityVSQuantity of substance

Solution Approach 1:

The patent changes the parameter of cooling redundancy assessment from binary (adequate/inadequate) to a quantitative measure based on capture indices and cooler contribution fractions. By computing the actual fractional contribution of each cooler to each rack's cooling and determining the minimum number of coolers whose failure would compromise redundancy, the system identifies the precise redundancy level provided, enabling cost-optimized provisioning rather than conservative over-provisioning.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP2656165B1Method for computing cooling redundancy at the rack level
Publication Date: 2018.03.28 SCHNEIDER ELECTRIC IT CORP
  • EP2656165B1 patent drawingFigure 1~2
  • EP2656165B1 patent drawingFigure 3
  • EP2656165B1 patent drawingFigure 4

AI summary

A system and method for designing aspects of a cooling system for a data center is provided. A method is provided including computing cooling redundancy at each rack position in a data center, in real time. The redundancy can be reported using the traditional N+1, N+2, etc. notation, where N is the number of coolers required to meet the primary cooling load type specification. The method substantially reduces the computational complexity of computing cooling redundancy.