Scribd
Upload
177 views2 pages

Quiz1 PDF

This document provides information about Quiz #1 for the class PE3105 Numerical Method. The quiz will be announced, closed book, and last 30 minutes. It is worth 50 points and covers topics from Week 1 on introduction to numerical methods and algorithms, and Week 2 on error and Taylor series. The quiz contains 3 problems - the first involves using Taylor series expansions to predict a function value, the second deals with error propagation in multiple variables, and the third involves applying Manning's formula for channel flow and determining which variable to measure more precisely based on a first order error analysis.

Uploaded by

yusufrevi
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
177 views2 pages

Quiz1 PDF

This document provides information about Quiz #1 for the class PE3105 Numerical Method. The quiz will be announced, closed book, and last 30 minutes. It is worth 50 points and covers topics from Week 1 on introduction to numerical methods and algorithms, and Week 2 on error and Taylor series. The quiz contains 3 problems - the first involves using Taylor series expansions to predict a function value, the second deals with error propagation in multiple variables, and the third involves applying Manning's formula for channel flow and determining which variable to measure more precisely based on a first order error analysis.

Uploaded by

yusufrevi
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
You are on page 1/ 2

Quiz #1

Type : Announced, Closed Book


Class : PE3105 Numerical Method
Topics : (Week 1) Introduction to numerical method and Algorithm
(Week 2) Error and Taylor Series
Date : 2 August 2019
Time : 30 minutes
Target : 50 Points

1. (Error and Taylor Series) Use zero- through third-order Taylor Series expansions to predict 𝑓(2)
for equation below
𝑓(𝑥) = 25𝑥 − 6𝑥 + 7𝑥 − 88
using a base point at 𝑥 = 1.
a. Write all derivatives (10 Points)
b. Compute the true percent relative error 𝜀 for each approximation. (10 Points)
c. In order to check its validity, sketch graph (cartesian coordinate) showing analytical and
numerical solutions for each approximation (10 Points)

2. (Error Propagation – Multiple Variables) Production rate for exponential decline is calculated by
𝑞=𝑞 𝑒
Calculate the rate at 𝑡 = 5 years if 𝑞 is estimated to be 830 stb/d ± 13% and the decline
exponent, 𝑎, is estimated to be 0.1 (fraction)/year ± 0.012 (fraction)/year. Indicate the relative
error in %. (20 Points)

3. (Error Propagation – Multiple Variables) Manning’s formula for a rectangular channel can be
written as
1 (𝐵𝐻) /
𝑄= 𝑆 /
𝑛 (𝐵 + 2𝐻) /
Where 𝑄 = flow (m3/s), 𝑛 = a roughness coefficient, 𝐵 = width (m), 𝐻 = depth (m), and 𝑆 =
slope. You are applying this formula to a stream where you know that the width = 20 m and the
depth = 0.3 m. Unfortunately, your know that the roughness is about 0.03 with range from 0.027
to 0.033 and the slope 0.0003 with a range 0.00027 to 0.00033. Use a first order error analysis to
determine the sensitivity of the flow prediction of these two factors. Which one should you
attempt to measure with more precision? (30 Points)

-END-
CHEAT SHEET

Taylor Series ---------------------------------------------------------------------------------------------------------------------------

Taylor Series Expansion

𝑓 (𝑥 )ℎ 𝑓 ( ) (𝑥 )ℎ 𝑓 ( ) (𝑥 )ℎ
𝑓(𝑥 ) = 𝑓(𝑥 ) + 𝑓 (𝑥 )ℎ + + + ⋯+ +𝑅
2! 3! 𝑛!
where ℎ = (𝑥 −𝑥 )

𝑓( ) (𝜉)ℎ
Remainder 𝑅 = or 𝑅 = 𝑂(ℎ )
(𝑛 + 1)!
Error Definitions ----------------------------------------------------------------------------------------------------------------------

True error 𝐸 = true value − approximation


true error
True percent relative error 𝜀 = 100%
true value

current approximation − previous approximation


Approximate percent relative error 𝜀 = 100%
current approximation

Stopping criteria |𝜀 | < 𝜀


where 𝜀 = (0.5 × 10 )% to assure that the result is correct to
at least 𝑛 significant figure; unless otherwise specified
Error Propagation --------------------------------------------------------------------------------------------------------------------

For 𝑛 independent variables 𝑥 , 𝑥 , … , 𝑥 approximated with 𝑥 , 𝑥 , … , 𝑥 having errors Δ𝑥 , Δ𝑥 , … , Δ𝑥 , the


error in the function 𝑓 can be estimated via
𝜕𝑓 𝜕𝑓 𝜕𝑓
Δ𝑓(𝑥 , 𝑥 , … , 𝑥 ) = Δ𝑥 + Δ𝑥 + ⋯ + Δ𝑥
𝜕𝑥 𝜕𝑥 𝜕𝑥

of 𝑓(𝑥) of 𝑥
Relative error Δ𝑓(𝑥 ) 𝑓(𝑥) − 𝑓(𝑥 ) Δ𝑥 𝑥 − 𝑥
or or
𝑓(𝑥 ) 𝑓(𝑥 ) 𝑥 𝑥

Provides a measure of the extent to which an uncertainty in 𝑥 is magnified by 𝑓(𝑥)


Condition number Relative error of 𝑓(𝑥)
Condition number =
Relative error of 𝑥

Numerical Differentiation – Finite Divided Difference ------------------------------------------------------------------------

𝑓(𝑥 ) − 𝑓(𝑥 )
First forward 𝑓 (𝑥 ) = + 𝑂(ℎ)
ℎ
𝑓(𝑥 ) − 𝑓(𝑥 )
First backward 𝑓 (𝑥 ) = + 𝑂(ℎ)
ℎ
𝑓(𝑥 ) − 𝑓(𝑥 )
First centered 𝑓 (𝑥 ) = − 𝑂(ℎ )
2ℎ

You might also like