The uncertainty of measurement characterizes the dispersion of the values that could reasonably be attributed to the stated value of the output quantity (measurand). In other words, the uncertainty U is the interval within which the conventional true value of the output quantity is likely to lie. For example, if Y is the calculated value of measurand (output quantity) from the data of measured input quantities and U is the uncertainty, then the conventionally true value is likely to lie between Y – U and Y + U.
The uncertainty of the output quantity Y consists of uncertainties in measurement of the input quantities. There are two methods of their evaluation, namely Type A evaluation and Type B evaluation. But in either case the quantities calculated are variances (Type A evaluation) or are in nature similar to the variances (Type B evaluation)
combined standard uncertainty
The estimated standard deviation of the estimate y is termed as combined standard uncertainty. It is denoted as uc.y/. The uncertainty uc.y is characterized by the positive square root of sum of the squares of the products of standard deviation and its corresponding partial derivative; similar terms are added for dependent input quantities for their covariances. The uc.y, in this case, is known as standard uncertainty.
For an independent quantity XP , there may be more than one source of uncertainty, a quantity similar to standard deviation is determined for each source and their squares are added. The square root of the sum is termed as standard uncertainty of XP and is denoted as u.xp/. Numerically the standard uncertainty is equal to the combined standard deviation of XP .
Source: measurement uncertainty
