Scribd
Upload
51 views19 pages

Measurement Techniques - Week 4

Uploaded by

fhggdhd
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
51 views19 pages

Measurement Techniques - Week 4

Uploaded by

fhggdhd
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/ 19

MEASUREMENT TECHNIQUES

Week 4
Uncertainty Analysis

• With this method developed by Kline and McClintock, a systematic


approach to the sensitivity of the measurement results is proposed and a
certain error range can be determined. The biggest advantage of this
method over other methods is that the variable that causes the most errors
among other variables can be easily detected.
• Thanks to this method, the variable that causes the most errors is easily
detected. Let the measured size of the system be R and the n independent
variables affecting this size are 𝑥1 , 𝑥2 , … , 𝑥𝑛 . In this case, the measured size
can be expressed mathematically as in the following equation.
𝑅 = 𝑅(𝑥1 , 𝑥2 , 𝑥3 , … , 𝑥𝑛 )
• If the error rate for each independent variable is Wi (W1, W2,…, Wn), then the
error rate WR of the R magnitude can be calculated with the help of the
following equation.
𝜕𝑅 2
𝜕𝑅 2
𝜕𝑅 2
𝜕𝑅 2 1ൗ
𝑊𝑅 = [( 𝑊1 ) + ( 𝑊2 ) + ( 𝑊3 ) + ⋯ + ( 𝑊𝑛 ) ] 2
𝜕𝑥1 𝜕𝑥2 𝜕𝑥3 𝜕𝑥𝑛
• Example: While calculating the mechanical power, the formula 𝑃 = 𝐹. 𝑣 is
used. It was obtained as 𝐹 = 780 𝑁 ± 10 𝑁 and 𝑣 = 0.61 𝑚/𝑠 ± 0.06 𝑚/𝑠. In this
case, calculate the error according to the uncertainty analysis.
• Example: A hole with a diameter d = 55 mm ± 0.2 mm is drilled into a cube
with a side a = 120 mm ± 0.3 mm.
a) Find the nominal volume of the object.
b) Find the error in the calculated volume by uncertainty analysis.
• Example: Calculate the nominal volume of the cylinder with a height h = 70
mm ± 0.1 mm and a diameter d = 15 mm ± 0.1 mm and the error in the
calculated volume according to the uncertainty analysis.
Normal Error Distribution
• Because the shape of the normal distribution function
looks like a bell, it is often referred to as the bell curve. The
distribution is expressed in mean ( 𝜇 ) and standard
deviation (𝜎) values and is defined in the range of -∞ and +
∞. The area under the bell curve gives the probability that a
certain error will recur. The 𝜎 shown on the x-axis is the
standard deviation of the measurement group. The normal
distribution curve created in the measurements made in a
small number of measurements helps to determine the
error probability of the result.
1 1 𝑥−𝜇
−(2)[ 𝜎 ]2
𝑓 𝑥 = 𝑒
𝜎 2𝜋
Standard Normal Distribution

• The special case where the mean is 0 (zero) and the


standard deviation is 1 (one) is called the standard
distribution. Standard normal distribution chart is
generally used in applications. The operations are
continued by transforming the x random variable in
the normal distribution calculations into z, which is
the variable of the standard normal distribution.
𝑥 − 𝜇
𝑧=
𝜎
Chauvenet Criterion

• Not all the measured values made can be completely correct.


The Chauvenet Criterion is used in order to obtain more
optimum results by eliminating the doubtful experimental
findings obtained as a result of the measurement. In order to
eliminate suspicious results, deviation values (𝑑𝑖 = 𝑥𝑖 − 𝑥)ҧ and
standard deviation (𝜎) are calculated.
• Example: The results of the measurements made in the laboratory
environment to determine a produced spring constant (k) are given
in a table. Evaluate the measurement results according to the
Chauvenet criteria.
Spring Constant
Number 𝒅𝒊 = 𝒙𝒊 − ഥ
𝒙 𝒅𝒊 /𝝈
(N/m), 𝒙𝒊
1 50,10

2 50,25

3 50,15

4 47,95

5 49,90

6 50,00

7 49,85

8 50,05

9 50,20

10 46,90
Least Squares Method

• The least squares method is a standard regression method used to write the
mathematical connection between two physical quantities that vary
depending on each other as an equation that is as realistic as possible. In
other words, this method serves to find a function curve that will pass "as
close as possible" to the data points obtained from the measurement result.
• 𝑦𝑖(𝑥): Experimental values
• 𝑦ഥ𝑖(𝑥) : Predicted values (regression values)
• 𝑞𝑖 : The difference between these two values
𝑛

𝑆 = ෍ 𝑞𝑖2 = 𝑆ℎ𝑜𝑢𝑙𝑑 𝑏𝑒 𝑚𝑖𝑛𝑖𝑚𝑢𝑚


𝑖=1
𝑦ത = 𝑎 𝑥𝑖 + 𝑏
𝑛 𝑛

𝑆 = ෍ 𝑞𝑖2 = ෍(𝑦𝑖 − 𝑎𝑥𝑖 − 𝑏)2


𝑖=1 𝑖=1

෍ 𝑦𝑖 = 𝑎 ෍ 𝑥𝑖 + 𝑛 𝑏

෍ 𝑥𝑖 𝑦𝑖 = 𝑎 ෍ 𝑥𝑖2 + 𝑏 ෍ 𝑥𝑖

𝑛 σ 𝑥𝑖 𝑦𝑖 − ( σ 𝑥𝑖 ) (σ 𝑦𝑖 )
𝑎=
𝑛 σ 𝑥𝑖2 − (σ 𝑥𝑖 )2
(σ 𝑦𝑖 )(σ 𝑥𝑖 )2 − (σ 𝑥𝑖 𝑦𝑖 ) ( σ 𝑥𝑖 )
𝑏=
𝑛 σ 𝑥𝑖2 − (σ 𝑥𝑖 )2
Correlation Coefficient

• The correlation coefficient (R) was defined as the indicator of how well the
curve obtained by the least squares method complies with the actual
measurement results. The correlation coefficient ranges from 0 (zero) to 1
(one). The zero value of R indicates that there is no agreement between the
curve obtained and the test results. Therefore, it is desired that the R value
be close to 1.
1ൗ
2
𝜎𝑦,𝑥 2
𝑅 = 1−
𝜎𝑦2
1ൗ
σ𝑛𝑖=1(𝑦𝑖
− ത 2
𝑦) 2
𝜎𝑦 =
𝑛−1
1ൗ
σ𝑛𝑖=1(𝑦𝑖
− 𝑦𝑖𝑘 )2 2
𝜎𝑦,𝑥 =
𝑛−2
Regression Analysis with Excel Program

• Many mathematical operations can be performed easily using the Excel


program. We can also perform regression analysis with the help of this
program.
• Example: The temperature-heat transfer coefficient change results of a
steel material are given in the table below. Find the appropriate equality
between temperature and heat conduction coefficient by performing
regression analysis in Excel program.

Coefficient of Thermal Conductivity Temperature


𝒚𝒊 = (𝑾/𝒎° 𝑪) 𝒙𝒊 (°𝑪)
73 5
65 90
64 205
55 297
50 407
41 503
40 608
Analysis of Experimental Findings
• While analyzing the experimental findings;
• It should examine the stability of the data.
• Where appropriate, statistical analysis methods should be applied to the
data.
• The results must be found uncertain.
• It is necessary to have information about the data to be obtained by
examining the theory.
• Data should be linked to relevant physical theories.

You might also like