11) Uncertainty

Question 1

Sources of uncertainty:

Answer 1

Incomplete definition of the test

Imperfect realisation of the test procedure

Inadequate understanding of environmental conditions

Sampling strategy not truly representative

Personal bias in reading analogue instruments

Question 2

What are the two categories of errors?

Answer 2

Random and systematic.

Question 3

What are the two groups of uncertainty components?

Answer 3

Type A and Type B

Type A - From a numerical data set - statistical calcs from series of repeated observation. Involves standard deviation calcs.

Type B - calculated by another other means than that of type A. Data from manufacturers specs, calibration certificates. Also use estimated standard deviations.

Question 4

Describe the stages of calculating uncertainty using type A method.

Answer 4

Data set -> calculate mean -> calculate sd -> apply factor -> combine to form CSU -> multiply by coverage factor k.

Question 5

What is the final stage of uncertainty analysis?

Answer 5

Combine components to achieve the combined standard uncertainty - CSU

Question 6

How do you calculate combined uncertainty?

Answer 6

CSU = sqrt(a^2+b^2+c^2)

Question 7

How do you convert the CSU to an expanded uncertainty?

Answer 7

Multiply by the coverage factor - K

U = k * Uc

Question 8

Type B example: A calibration certificate states the measurement uncertainty as +/- 0.1% at a 95% confidence level.

Answer 8

u(x) = +/- (0.1 / 2)% = +/- 0.05% of the reading

Question 9

For a +/- 1% type B evaluation, divide by sqrt(3). For example for an instrument with a stated accuracy of +/- 1%:

Answer 9

u(x) = +/- 1 / sqrt 3 = +/- 0.58% of the reading