無扇式LED背光板之熱計算-使用有限元素法

Abstract

本論文建立一組實驗用以評估一無扇式(也即意味著無雜訊及無功率之散熱處理法) LED背光板的散熱問題，並用傳統熱傳理論瞭解FEM法模擬數據及實驗數據的差異原因。於在第一章中我們將概括討論現階段LED散熱技術，在第二章我們將推導在工程計算上的熱傳理論。熱傳公式及一些無單位數將被定義，並以無扇式LED背光板的應用來推導。邊界層理論，等溫、等熱流垂直板，等熱流非對稱之垂直兩平行板自由對流以及最佳及最大氣渠板距將被討論。 現今，藉由快速的計算機科技，使用電腦工具於有限元素法(FEM)的計算不再昂貴，實際的有限元素問題可能包含數萬甚至數百萬個元素及節點，因此常用現有的商業化套裝軟體來求解。簡要的有限元素法用於熱傳問題將於第三章中討論。一個用基本矩形元素溫度節點的手算勁度矩陣將被推導並與套裝軟體內建之相似元素做比較。 在第四章中，我們將建立一組實驗來量測並模擬在LED-背光板板上的溫度分佈，單層板及氣渠結構的背光板將被建模。 溫度相關之局部熱傳係數(半高處)在不同面及不同功率的值將被估算，並據以代入模擬模組，我們發現這樣的做法，實驗數據與模擬數據可以相當的接近，從低(100W)到高功率(400W)層流範圍及不同結構應用，另外，板的下方為自由對流流體的入口處，這個區域可被預期有較薄的邊界層厚度及較大的局部熱傳係數，然而在模擬過程並未被考慮，以至於觀察到實驗數據與模擬數據有稍大的誤差(仍在20%以內)。 於第五章中，我將對這些研究做總評及後續工作。

The thesis setup a set of experiment to investigate the heat dissipation in a fan-less (which also refer to noise-free and power-free thermal management method) LED Back-lighting Unit (LED-BLU). The reason of deviation between FEM simulation data and experimental data will be understood by traditional heat transfer theorem. In chapter 1, we will generally discuss current LED thermal dissipation techniques. In chapter 2, heat transfer equations and some of dimensionless numbers will be defined and derived for LED-BLU application. Boundary layer theorem, isothermal and isoflux vertical plates, asymmetric isoflux vertical parallel plates of free convection, optimum and maximum channel space will be discussed. Nowadays, with rapidly advancing computer technology, use of the computer as a tool in the Finite element method (FEM) is no more dispensable. Realistic finite element models might consist of up to tens of thousands, or even several millions of elements,nodes and therefore usually solved by using commercial software packages. A briefly finite element method using in heat transfer problem will be discussed in chapter 3. A hand calculated stiffness matrix by using basic rectangular element with nodal temperature will be derived and compared with similar built-in element of software package. In chapter 4 , we will setup a experiment to measure and to simulate the temperature distribution on the LED-BLU. The single sheet and air-duct structure BLU will be modeled. The temperature dependent local heat transfer coefficient at half height will be estimated for different surfaces and different powers, these values will then substitute into simulation model. By this way, we found most of experiment data can be very close to simulated data from lower (100W) to higher (400W) power in laminar flow range and in different structures application. On the other hand, the bottom area of the plate is considered as the entry area of free convection flow, so the thinner boundary layer thickness and higher local heat transfer coefficient can be expected in these area. However, it is not be considered and result in larger deviation was observed (within 20%). In chapter 5 , I will make an overall comment on these study and the follow-up.