Scribd
Upload
3 views2 pages

APL 321 Lab1

This report analyzes the numerical solution of the 1-D convection-diffusion equation with a Peclet number of 50 and boundary conditions ϕ(0) = 0 and ϕ(1) = 1. It compares first-order upwind differencing and second-order central differencing schemes, noting stability and oscillation issues, and evaluates discretization error and order of accuracy. The findings confirm first-order accuracy for UDS and second-order accuracy for CDS.

Uploaded by

Pratham Chugh
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
3 views2 pages

APL 321 Lab1

This report analyzes the numerical solution of the 1-D convection-diffusion equation with a Peclet number of 50 and boundary conditions ϕ(0) = 0 and ϕ(1) = 1. It compares first-order upwind differencing and second-order central differencing schemes, noting stability and oscillation issues, and evaluates discretization error and order of accuracy. The findings confirm first-order accuracy for UDS and second-order accuracy for CDS.

Uploaded by

Pratham Chugh
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

APL321: Lab Assignment 1

Error Analysis of 1D Convection-Diffusion Equation


Pratham Chugh
January 25, 2026

1 Introduction
This report analyzes the numerical solution of the non-dimensional 1-D convection-diffusion
equation for a scalar ϕ(x):
dϕ d2 ϕ
Pe = 2 (1)
dx dx
The Peclet number is chosen as P e = 50, with boundary conditions ϕ(0) = 0 and ϕ(1) = 1.
The analytical solution used for comparison is:

exp(xP e) − 1
ϕexact (x) = (2)
exp(P e) − 1

2 Question 1: Numerical Solution and Comparison


2.1 Discretization Schemes
The domain is discretized into N points with uniform spacing h = 1/(N − 1). We express the
discrete form as AP ϕi + AW ϕi−1 + AE ϕi+1 = 0.

2.1.1 Part (a): First-Order Upwind Differencing Scheme (UDS)


For P e > 0, information is taken from the upwind node (i − 1):

• AW = 1
∆x2
+ Pe
∆x

• AE = 1
∆x2

• AP = − ∆x2 2 − Pe
∆x

2.1.2 Part (b): Second-Order Central Differencing Scheme (CDS)


Both convective and diffusive terms utilize central differences:

• AW = 1
∆x2
+ Pe
2∆x

• AE = 1
∆x2
− Pe
2∆x

• AP = − ∆x2 2

1
2.2 Results and Observations
Numerical solutions were generated for N = 11, 21, 41, 81.

• Stability: UDS remained stable across all N , though it exhibited numerical diffusion.

• Oscillations: CDS produced non-physical oscillations at N = 11 and N = 21 because


the cell Peclet number exceeded the stability limit (P e · ∆x > 2).

3 Question 2: Discretization Error Analysis


3.1 Error Definition
Discretization error ϵh is defined using a standard deviation norm:
v
u N
uX (ϕh (xi ) − ϕexact (xi ))2
ϵh = ||ϕh − ϕexact || = t (3)
i=1
N

where ϕh is the discrete solution and ϕexact is the analytical solution.

3.2 Convergence Results (Log-Log Plot)


The error ϵh was plotted against h on a log-log scale for N = 11, 21, 41, 81. Reference lines
y1 = c1 h and y2 = c2 h2 were included to visualize the order of accuracy.

3.3 Commentary on Order of Accuracy


• UDS (Part a): The error curve aligns with the O(h) slope, confirming first-order accu-
racy.

• CDS (Part b): The error curve aligns with the O(h2 ) slope, confirming second-order
accuracy.

4 Question 3: Local Order of Accuracy


Without the exact solution, the order mh is calculated using three grid levels:
 
1 ||ϕ2h − ϕ4h ||
mh = ln (4)
ln 2 ||ϕh − ϕ2h ||

As h → 0, mh for UDS approaches 1 and CDS approaches 2.

You might also like