臺灣博碩士論文加值系統

對於兩相紊流，在理論上最常使用的是Eulerian-Lagrangian 的方法，所
謂 的Eulerian-Lagrangian的方法是指在連續相的控制方程式是以歐氏
法(Eulerian)表示, 包含了質量、動量守恆定律，以及氣相方面的紊流動
能方程式及紊流動能消散率方程式，而在分散相也就是顆粒方面的運動方
程式則是以拉氏法(Lagrangian)表示; Eulerian -Lagrangian 的方法通
常有兩種模式 , 一種是 DSF( deterministic separated flow)模式,但
並沒有考慮紊流的分散效應, 另一種是 SSF (stochastic separated
flow) 模式 , 粒子所受的紊流分散效應 ( turbulence dispersion
effect)以蒙地卡羅法來處理； 早期使用的是DSF模式,近期則使用SSF模
式,但是使用SSF模式 , 達到統計上穩定 (statistically stationary)的
解時, 所需的代表性計算顆粒數目非常多, 而且以往針對顆粒的阻力係數
只考慮假設為穩態的阻力係數使得其顆粒紊流動能及大顆粒的擾動速度的
機率密度分佈函數的數值計算結果與實驗值比較有相當的差距; 本研究針
對上述缺點提出一改良的模式, 即由DSF模式加上解顆粒紊流動能的方程
式,吾人簡稱為DM模式;且針對顆粒的阻力係數, 加入考慮暫態(unsteady)
的阻力係數,使得所解得顆粒的紊流動能及顆粒擾動速度的機率密度分佈
函數能與實驗值更接近。本研究同時針對實驗量測數據與數值計算兩者之
間的比較基準進行探討；實驗量測在每一截面所量得的通量小於該截面由
數值計算所得的最小通量,則該點的實驗量測數據即可刪去,如此可使得比
較的基準相同,同時亦針對所給予不同數目的代表性顆粒,由最小通量的變
化 ,當其趨於固定值時,則可視此改良模式達到統計上穩定的解時所需最
少的代表性計算顆粒數目。

Eulerian-Lagrangian method is a common used method for two
phase turbulent flow in engineering application. Eulerian-
Lagrangian method means that the governing equation of the
continuous phase are formulated in an Eulerian frame including
mass,momentum conservation laws and turbulent closure modeling
equations.Particle equations are solved by a stochastic
discrete particle technique based on Lagrangian formulation
since most of the test problems are of polydispersed flows.
There are two models widely used for the Eulerian-Lagrangian
method,one is deterministic separated flow (DSF) model which
does not take into account the effect of turbulence dispersion,
the other is stochastic separated flow(SSF)model. For SSF model,
the turbulent dispersion effect are considered by using the
Monte-Carlo method. There should have many representive,
computational drops for dispersed phase to obtain a
statistically stationary solution for SSF model. Also the
steady state formula of the drag coefficient is employed in the
formulation of the Lagrangian equation of motion before. By
using SSF model and the steady state of drag coefficient, it
introduces some errors in the numerical solution for drop
tuebulent kinetic energy and the pdf of velocity fluctuation of
large drops. To overcome the above deficiency, a novel model by
using DSF model and solving the equation of turbulent kinetic
energy,which is called DM model,is proposed. The unsteady drag
coefficient of drop is used to improve the predict particle and
the pdf of velocity fluctuation. A selection criterion for
measured data is also proposed as follows. When the measured
mass flux at each point of one section is lower than the
minimum flux by numerical computation, these measured data
should not be included for comparison. When the minimum flux
value approaches a stable value as increasing the number of
representive computational drops,we regard it reaching the
statistically stationary solution.