本研究係設計一套應用於網路伺服器系統之智慧型風扇節能系統。網路伺服器係屬於全天候工作的產品,本身於作業時亦會有散熱的需求,一般在系統設計上均會採用三到四組的散熱風扇來幫助系統散熱,因此,如果可以針對此散熱系統做節能上的控制,則長久下來必可節省相當可觀的能源耗損。 研究中利用線性搜尋法來依序控制風扇轉速,以求得系統中散熱片的散熱飽和點,並依此飽和點來當作風扇轉速的控制目標,以求得在相同的散熱功效下兼具省能的最佳風扇轉速控制點。在研究過程中,透過迴歸分析發現風扇轉速與溫度變化的線性關係,依此關係的成立進而設計出二元搜尋系統來加速最佳風扇控制值的求得,大幅改善線性搜尋法求解過慢的缺點。 本論文主要貢獻為利用自行設計之智慧型風扇控制系統,加以應用在網路伺服器上,透過實際實驗證明,除了可以將系統溫度控制在一穩定的溫度區間外,更可以有效的節省不必要的風扇能源耗損。
This research is to develop an effective control strategy for the fans in a LAN (Local Area Network) server system. The LAN server system is a full time working product which in most cases requires a heat dissipation system. Typically, multiple sets of fan are required to help server system dissipate heat, which implies that the energy consumed by these fans should not be overlooked while the issue of power saving is extensively addressed in different fields nowadays. In this thesis, we introduce an effective control strategy such that the power consumed by the heat dissipation system is significantly reduced. Two stage fan control method is applied on most of the LAN server system for dissipate heat nowadays, and this method simply controls the fans for power saving in a rough manner. Thus, we try to find out the relation between fan speed and system temperature by using regression analysis method. And then, according to the speed-temperature relation, we develop an effective control strategy with linear search and binary search method to apply on LAN server system. The experimental results reveal that, with our control strategy, not only the system temperature is kept in a stable temperature range, but also energy consumption of the fans is appreciably reduced.