Browse Prior Art Database

Liquid Flow Through Heat Exhanger with Multiple Coolant Channels

IP.com Disclosure Number: IPCOM000201589D
Publication Date: 2010-Nov-15
Document File: 3 page(s) / 44K

Publishing Venue

The IP.com Prior Art Database

Abstract

A supercomputer cooling system with at least two separate coolant channels, one at ambient temperature for economical operation and the other at a lower temperature for maximum cooling performance when the supercomputer is under maximum computational load. This permits the nominal channel of the coolant system to be designed for nominal computational loads, while the auxiliary channels is designed to handle peak loads. It is contended that this separation of coolant system performance based on computational loading will result in economies of HW and operation.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 01 of 3

Liquid Flow Through Heat Exhanger with Multiple Coolant Channels

Just as a 2500 pound car w/500 HP engine consumes more gas than the same car w/150 HP engine, even though the city driving is exactly the same, a supercomputer's coolant system must be designed and built to cover peak demands rather than average operational conditions even if peak loads represent only a small faction of system operational conditions. This means that a supercomputer's coolant system may somewhat resemble a 500 HP engine when a 150 HP engine will do. Implying that improved efficiencies and better Green practices would be possible if a new supercomputer coolant system could be developed which would adjust from nominal to peak loading conditions w/o loss of performance or efficiency/economy.

In addition, even though most supercomputer systems have coolant system which scale in performance based on system workload (power being dissipated), there is a very significant time-lag between device TJ 's rising and driving greater coolant flow and the eventual cool down of the system. In fact, there could be many minutes between initial TJ rise and the eventual cool down to the increased coolant flow response. The acceleration factor, approximately doubling for each 10C rise for 1st order rate kinetic, implies that the lag-time is consuming 2x or higher device steady-state life for every minute of time-lag (ie, significantly reducing system reliability for every work-load induced thermal excursion).

The invention is a new method and means to permit the design and build of a computer's liquid cooling system to maximize efficiency and Green performance while also fully accommodating the host system variable cooling demands. This is done by separating the nominal (steady-state) operational load conditions, which determine the nominal/steady-state cooling requirements, from the peak system load conditions, which determines the Peak-Demand cooling requirements. Thus the invention is an On-Demand liquid flow through (LFT) coolant method and means. The valve type depicted in the figures of this disclosure are crude depictions of common / known valve means and are not claimed.

The two principle aspects of the invention are: 1) deign and build a nominal/steady-state coolant channel/system; 2) design and build of an embedded Peak-Demand coolant channel/system; and 3) the integ...