ISO 6974-1-2012 pdf free download

ISO 6974-1-2012 pdf free download.Natural gas — Determination of composition and associated uncertainty by gas chromatography — Part 1:General guidelines and calculation of composition
Gaz naturel — Determination de Ia composition et de Iincertitude associée par chromatographie en phase gazeuse — Partie 1: Lignes directrices générales et calcul de Ia composition.
The CRMs selected may be multi-component or binary mixtures with appropriate uncertainty and should always be fit for purpose (see ISO 6143 and ISO 6142[J), i.e. the uncertainty of each component should be small enough that it gives an acceptably small final uncertainty after propagation through the full uncertainty budget.
NOTE It is not possible to make a mixture with every component at the highest level required, or at the lowest, so most of the multi-component CRMs will have compositions that are different from rear natural gas mixtures. Provided the mixtures are shown to be stable in storage and use, this will not create a problem.
CRMs are generally provided with expanded uncertainties calculated with a coverage factor. k, normally equal to 2. A coverage factor of 2 should therefore be used as the default option in the absence of any other evidence.
6.5.4 MeasurIng reference gases
Analyse each of the CRMs. It is recommended that a minimum of 10 analyses be performed for each reference gas so as to ensure that the mean response data and their standard uncertainties are determined with a precision that is fit for purpose.
NOTE The uncertainties referred to in this subclause are not used in this part of Iso 6974. but are required if ISO 6974-2 is to be used to determine uncertainties of mole fractions.
Depending on the number of repeated measurements, the uncertainty of the uncertainty of a mean value
(i.e. the relative standard deviation of the standard deviation of a mean value) can be surprisingly large: for
10 measurements it is 24 % (see ISO/fEC Guide 98-3:2008, E.4.3). It is therefore not recommended to use a
smaller number of repeated measurements when determining the standard deviation of a mean value.
Tabulate the individual responses to each component in each standard for each replicate. The data may be inspected for outliers by using a suitable outlier test (see Annex E). If outliers are found, the data should be investigated; outliers should be rejected only if there are sound technical reasons for doing so.
6.5.5 Regression analysis
6.5.5.1 Introduction to regression analysis
It is strongly recommended that generalized least squares (GLS) analysis be undertaken. GLS analysis, which complies with ISO 6143. requires that the uncertainties of the input data be determined and performs regression weighted to these uncertainties, The value of each component mole fraction is calculated, along with its uncertainty. GLS shall therefore be used if ISO 6974-2 will subsequently be used to determine the uncertainty of mole fractions.
For cases where the uncertainties of component mole fractions (and therefore compliance with ISO 6974-2) are not required, either GLS analysis with zero (or negligibly small) input uncertainties or ordinary least squares (OLS) analysis may be used. However, even when the uncertainties of component mole fractions are not required, it is recommended that GLS be used with defined input uncertainties whenever possible. This gives the most reliable determination of component mole fractions.ISO 6974-1-2012 pdf free download.