Measuring systems Problem set n 10

Measuring systems
Problem set 10
Data analysis

Exercise 1 (Estimate of the average)

We can begin by expressing the number of samples n that will be recorded for each measurement and the
standard deviation of the population (standard measurement for a single measurement) , considering
the known reference velocity :
= = 200 and =

a) = = 4.95 /
b) If we quadruple the number of measurements N, we find the standard deviation 4 for the calculated
average:
,4 = ,

The standard deviation of the average depends on the number of points used to calculate the average
itself. It has nothing to do with how often we repeat this.

c) If the acquisition time is 2 , for a standard deviation ,2 of the averages:

2

= = 2 2 =
= 5.44

d) The standard deviation of the averages ,3 , for a sampling frequency which is two times smaller
and the acquisition period of , will be:

=
2
= = 2 = 0.5 /

Exercise 2 (Types of error, confidence level and number of measurements)

a) The accuracy of is expressed as the systematic error . Since an error on the central tendency
is described as a systematic error, it must be sufficiently far from the central tendency (e.g.
> 3 for a confidence level of 99%. This is a somewhat arbitrary number and you could have
chosen another confidence level but you need to declare the confidence level associated with your
choice of z p).
= = 0.6

Fidelity (precision) is expressed as the statistical error according to the chosen confidence level for
. associated with 0 and 1 :
,0 = = 0.15 ,1 = 2 = 0.3

The statistical error ,2 for 2 is calculated according to the table of the normal distribution:
12
= = 0.05
2
= 1.646
,2 = = 0.25
Measuring systems Problem set n 10

The total error of corresponds to the sum of and :

,0 = 0.85 ,1 = 1.0 ,2 = 0.95

b) The statistical error ,2 on the average for 2 , knowing that = 200 :

,2
,2 = = 0.0175

2
c) = 2 609