erf(0)=0
erf(0.025)=0.028204
erf(0.05)=0.056372
erf(0.075)=0.08447
erf(0.1)=0.11246
erf(0.125)=0.14032
erf(0.15)=0.168
erf(0.175)=0.019547
erf(0.2)=0.227
erf(0.225)=0.24967
erf(0.25)=0.27633
erf(0.275)=0.30266
erf(0.3)=0.32863
erf(0.325)=0.35421
erf(0.35)=0.37938
erf(0.375)=0.40412
erf(0.4)=0.42839
erf(0.425)=0.45219
erf(0.45)=0.47548
erf(0.475)=0.49826
erf(0.5)=0.5205
erf(0.525)=0.54219
erf(0.55)=0.56332
erf(0.575)=0.58388
erf(0.60)=0.60386
erf(0.625)=0.62324
erf(0.65)=0.64203
erf(0.675)=0.66022
erf(0.70)=0.6778
erf(0.725)=0.69478
erf(0.75)=0.71116
erf(0.775)=0.72693
erf(0.8)=0.7421
erf(0.825)=0.75668
erf(0.85)=0.77067
erf(0.875)=0.78407
erf(0.9)=0.79691
erf(0.925)=0.80918
erf(0.95)=0.82089
erf(0.975)=0.83206
erf(1.0)=0.8427
erf(1.05)=0.86244
erf(1.10)=0.8802
erf(1.15)=0.89612
erf(1.20)=0.91031
erf(1.25)=0.9229
erf(1.30)=0.93401
erf(1.35)=0.94376
erf(1.40)=0.95229
erf(1.45)=0.9597
erf(1.50)=0.96611
erf(1.55)=0.97162
erf(1.60)=0.97635
erf(1.65)=0.98038
erf(1.70)=0.98379
erf(1.75)=0.98667
erf(1.80)=0.98909
erf(1.85)=0.99111
erf(1.90)=0.99279
erf(1.95)=0.99418
erf(2.0)=0.99432
Cooling of a dike: Analytical vs Numerical solutions...
Exercise 5: The analytical solution that describes the heat transfer through time across of a dike 2w meters wide is:
erf(x) is the error function. It was first tabulated in the mid-1800's during the development of probability theory. Its properties are:
erf(0)=0
erf(∞)=1
erf(-x)=-erf(x)
The complementary error function erfc(x) is defined as: erfc(x)=1-erf(x)
An approximation of the error function is:
0≤x<0.6: erf(x)=x
0.6<x≤1.2: erf(x)=0.84+log(x)
1.2<x≤2: erf(x)=0.9+0.35.log(x)
x>2: erf(x)=1
with erf(-x)=-erf(x)
• At the time of emplacement the material inside a 20 meter wide dike was at T1=900ºC whereas the surrounding host rock was T0=300ºC. Using the analytical solution above, calculate the temperature profile across the dike at a time t=1month, 1 year, 10 years after emplacement. (Diffusivity: 10-6 m2.s-1)
• Build an Ellipsis model using a 1x6 box, level:4, temperature-independent viscosity (η 0=2500). The dike, 0.2 thick, is centered at x=3. Background temperature: 300ºC, temperature in the dike: 900ºC. Since lengths are divided by 100 the diffusivity must be scaled to maintain the representativity of the model. This scaling can be done by noting that the dimension of diffusivity is m2.s-1. Since the scaling factor for length (m) is 10-2 that of surface (m2) is (10-2)2. The analytical solution tells us that T(x, t) is depends on x/(2.√ (λ .t)). The temperature is not affected by reducing the numerator as long and the denominator is reduced by the same proportion. Compare the analytical and numerical solutions.