1、BSI Standards PublicationBS EN 15948:2015Cereals Determinationof moisture and protein Method using Near-Infrared-Spectroscopy in whole kernelsBS EN 15948:2015 BRITISH STANDARDNational forewordThis British Standard is the UK implementation of EN 15948:2015. Itsupersedes BS EN 15948:2012 which is with
2、drawn.The UK participation in its preparation was entrusted to TechnicalCommittee AW/4, Cereals and pulses.A list of organizations represented on this committee can beobtained on request to its secretary.This publication does not purport to include all the necessaryprovisions of a contract. Users ar
3、e responsible for its correctapplication. The British Standards Institution 2015. Published by BSI StandardsLimited 2015ISBN 978 0 580 86711 8ICS 67.060Compliance with a British Standard cannot confer immunity fromlegal obligations.This British Standard was published under the authority of theStanda
4、rds Policy and Strategy Committee on 30 April 2015.Amendments issued since publicationDate Text affectedBS EN 15948:2015EUROPEAN STANDARD NORME EUROPENNE EUROPISCHE NORM EN 15948 March 2015 ICS 67.060 Supersedes EN 15948:2012English Version Cereals - Determination of moisture and protein - Method us
5、ing Near-Infrared-Spectroscopy in whole kernels Crales - Dtermination de la teneur en eau et en protines - Mthode utilisant la spectroscopie dans le proche infrarouge sur des grains entiers Getreide - Bestimmung der Feuchte und des Proteins - Verfahren der Nahinfrarot-Spektroskopie bei ganzen Krnern
6、 This European Standard was approved by CEN on 5 January 2015. CEN members are bound to comply with the CEN/CENELEC Internal Regulations which stipulate the conditions for giving this European Standard the status of a national standard without any alteration. Up-to-date lists and bibliographical ref
7、erences concerning such national standards may be obtained on application to the CEN-CENELEC Management Centre or to any CEN member. This European Standard exists in three official versions (English, French, German). A version in any other language made by translation under the responsibility of a C
8、EN member into its own language and notified to the CEN-CENELEC Management Centre has the same status as the official versions. CEN members are the national standards bodies of Austria, Belgium, Bulgaria, Croatia, Cyprus, Czech Republic, Denmark, Estonia, Finland, Former Yugoslav Republic of Macedon
9、ia, France, Germany, Greece, Hungary, Iceland, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, Netherlands, Norway, Poland, Portugal, Romania, Slovakia, Slovenia, Spain, Sweden, Switzerland, Turkey and United Kingdom. EUROPEAN COMMITTEE FOR STANDARDIZATION COMIT EUROPEN DE NORMALISATION EUROPI
10、SCHES KOMITEE FR NORMUNG CEN-CENELEC Management Centre: Avenue Marnix 17, B-1000 Brussels 2015 CEN All rights of exploitation in any form and by any means reserved worldwide for CEN national Members. Ref. No. EN 15948:2015 EBS EN 15948:2015EN 15948:2015 (E) 2 Contents Page Foreword 4 1 Scope 5 2 Nor
11、mative references 5 3 Terms and definitions .5 4 Principle 5 5 Method of analysis.5 5.1 General 5 5.2 Near Infrared Instrument .6 5.3 Prediction models 6 5.4 Initial validation of the model .6 5.4.1 General 6 5.4.2 Initial validation sample set 6 5.4.3 Initial validation performances .7 5.5 Update o
12、f calibration model and validation of new model 7 6 Sampling .7 7 Procedure .7 7.1 Preparation of the test sample .7 7.2 Measurement 7 7.3 Local validation of the method .7 7.4 Periodical adjustment of the instrument .8 7.5 Checking instrument stability 8 7.6 Follow up of method performance .8 8 Cal
13、culation and expression of results .8 9 Accuracy and precision of the method .9 9.1 Accuracy .9 9.2 Precision .9 9.2.1 General 9 9.2.2 Repeatability .9 9.2.3 Reproducibility .9 9.2.4 Critical difference 10 10 Test Report 11 Annex A (informative) Results of examples of interlaboratory test 12 A.1 FOS
14、S interlaboratory test 12 A.2 PERTEN interlaboratory test . 17 A.3 CHOPIN Technologies interlaboratory test 21 Annex B (informative) Validation of ANN prediction model WB003034 . 29 B.1 ANN prediction model WB003034 with associated database 29 B.2 Results of validation according to EN ISO 12099 29 B
15、.3 Stability and robustness 31 Annex C (informative) Validation of Inframatic 9500 prediction models 33 C.1 Perten Inframatic 9500 prediction models . 33 C.2 Results of validations according to EN ISO 12099 33 BS EN 15948:2015EN 15948:2015 (E) 3 Annex D (informative) Validation of Infraneo predictio
16、n models . 39 D.1 CHOPIN Technologies Infraneo prediction models . 39 D.2 Results of validation according to EN ISO 12099 39 Bibliography 45 BS EN 15948:2015EN 15948:2015 (E) 4 Foreword This document (EN 15948:2015) has been prepared by Technical Committee CEN/TC 338 “Cereal and cereal products”, th
17、e secretariat of which is held by AFNOR. This European Standard shall be given the status of a national standard, either by publication of an identical text or by endorsement, at the latest by October 2015, and conflicting national standards shall be withdrawn at the latest by October 2015. Attentio
18、n is drawn to the possibility that some of the elements of this document may be the subject of patent rights. CEN and/or CENELEC shall not be held responsible for identifying any or all such patent rights. This document supersedes EN 15948:2012. The following modifications were made in this new edit
19、ion: Annexes have been enhanced with information on 2 other NIR instruments based on the results of the interlaboratory tests and the models of prediction. According to the CEN-CENELEC Internal Regulations, the national standards organizations of the following countries are bound to implement this E
20、uropean Standard: Austria, Belgium, Bulgaria, Croatia, Cyprus, Czech Republic, Denmark, Estonia, Finland, Former Yugoslav Republic of Macedonia, France, Germany, Greece, Hungary, Iceland, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, Netherlands, Norway, Poland, Portugal, Romania, Slovakia,
21、Slovenia, Spain, Sweden, Switzerland, Turkey and the United Kingdom. BS EN 15948:2015EN 15948:2015 (E) 5 1 Scope This European Standard defines a routine method for the determination of moisture and protein in whole kernels of barley and wheat using a near-infrared spectrophotometer in the constitue
22、nt ranges: a) for wheat: 1) moisture content minimum range from 8 % to 22 %; 2) protein content minimum range from 7 % to 20 %. b) for barley: 1) moisture content minimum range from 8 % to 22 %; 2) protein content minimum range from 7 % to 16 %. This European Standard describes the modalities to be
23、implemented by the supplier (5.3 and 5.4) and the user of the method. 2 Normative references The following documents, in whole or in part, are normatively referenced in this document and are indispensable for its application. For dated references, only the edition cited applies. For undated referenc
24、es, the latest edition of the referenced document (including any amendments) applies. EN ISO 12099:2010, Animal feeding stuffs, cereals and milled cereal products Guidelines for the application of near infrared spectrometry (ISO 12099:2010) ISO 5725-2, Accuracy (trueness and precision) of measuremen
25、t methods and results Part 2: Basic method for the determination of repeatability and reproducibility of a standard measurement method 3 Terms and definitions For the purposes of this document, the terms and definitions given in EN ISO 12099:2010 apply. 4 Principle The method is based on Near-Infrar
26、ed (NIR) spectroscopy, an indirect, correlative technique to predict the concentration of various constituents in organic samples. Linear or non-linear regression modelling is used to relate NIR spectra to moisture or protein concentrations determined by officially approved standard methods (e.g. ar
27、tificial neural network - ANN, Partial Least Square Regression - PLS). 5 Method of analysis 5.1 General According to this document, the method of analysis is defined as the association between a NIR instrument and a model of prediction. BS EN 15948:2015EN 15948:2015 (E) 6 5.2 Near Infrared Instrumen
28、t Based on diffuse reflectance or transmittance measurement covering the wavelength region of 700 nm to 2 500 nm or segments of this or at selected wavelengths. 5.3 Prediction models Each model for the prediction of protein and moisture contents in whole grain of wheat and barley is amongst others d
29、efined by: the number of samples used for the calibration development; the constituent ranges covered in the model for moisture and protein; the temperature range of the samples; the number and performance of involved reference labs; the stability of the model i.e. by number of harvests covered; the
30、 calibration file defined by its name and its IT name (for example CHECKSUM) insuring its integrity; the seasonal, geographic and genetic variations covered. 5.4 Initial validation of the model 5.4.1 General Since NIR analysis is an indirect, correlative technique, the results shall be validated aga
31、inst chemical analysis reference methods. It is important that the reference methods used are officially approved such as the methods described in the standards previously cited (Clause 2). The purpose of validation is to determine the root mean square error of prediction which depends at the same t
32、ime on the correlation, the bias and the slope. The root mean square error between chemical analysis methods and predictions shall be compared to calibration performance specifications and/or historical performance. 5.4.2 Initial validation sample set The initial validation of a calibration model sh
33、all be done in accordance with EN ISO 12099 using independent test sets of wheat and barley samples, originating from different countries and analyzed by the reference methods given in Clause 2. Requirements for the validation sample set are: at least 200 samples coming from 10 countries (20 represe
34、ntative samples min/country) distributed homogeneously over the entire constituent range; the part of the range without any reference sample shall not exceed 0,3 %; different scans from one sample shall not be considered as different samples; seasonal effects over at least a three year period, tempe
35、rature effects, instrument variation and the variability of reference data shall be included in the set. BS EN 15948:2015EN 15948:2015 (E) 7 5.4.3 Initial validation performances The results of the initial validation shall at least fulfil the specifications given in Table 1. Table 1 NIR performances
36、 for the determination of moisture and protein (see also Annex B) Moisture Wheat and barley Protein Wheat Protein Barley Overall accuracy expressed as SEP as constituent % w/w 0,24 % 0,27 % 0,27 % Constituent concentration in the independent validation data set Min 8 % 7 % d.m. 7 % d.m. Max 22 % 20
37、% d.m. 16 % d.m. NOTE The minimum performance given in Table 1 includes the variation of reference data as documented by the number of reference labs involved, regional and genetic variations, the number of countries and crop species involved and the robustness over the last five years (see also Ann
38、ex B). 5.5 Update of calibration model and validation of new model The prediction model in accordance with this standard shall be updated by the one issuing the calibration model to ensure inclusion of new climatic crop conditions and new varieties introduced on the market. These updates shall be ma
39、de by keeping the original database with addition of the new samples as needed. The new prediction model shall be updated according to EN ISO 12099. Validation shall be made according to the initial validation (5.4) and include at least 20 new samples. 6 Sampling Sampling is not part of the method s
40、pecified in this European Standard. A recommended sampling procedure is given in EN ISO 24333 5. It is important that the sample analyzed in routine is truly representative for the batch and has not been damaged or modified. 7 Procedure 7.1 Preparation of the test sample No specific sample preparati
41、on is required. 7.2 Measurement Follow the instructions of the instrument manufacturer. 7.3 Local validation of the method Before use, the method shall be validated on an independent test set that is representative of the sample population to be analyzed. For the determination of bias, at least 10 s
42、amples are needed; for the determination of Standard Error of Prediction (SEP, see EN ISO 12099:2010, 6.5) at least 20 samples are BS EN 15948:2015EN 15948:2015 (E) 8 needed. Validation shall be carried out for each sample type, constituent/ parameter and temperature (see EN ISO 12099:2010, 5.4). Bi
43、as or inherent systematic error, as described in EN ISO 12099:2010, Clause 6, is exhibited when the predicted results of a specific sample group or product show a mean offset value when compared to their reference values. This may occur with unique sample types. The bias (i.e. mean difference betwee
44、n the chemical analysis results and the predicted results) may or may not be statistically significant. Based on the procedure described in EN ISO 12099, a bias confidence limit may be calculated. When this limit is exceeded, a bias is implemented in the instrument software and the validation proces
45、s repeated. Refer to the manufacturer instructions and to EN ISO 12099 for procedure. 7.4 Periodical adjustment of the instrument To ensure its accuracy, each instrument shall be checked at least annually, against the reference method, either directly or through a master instrument. The execution of
46、 this check shall be performed on samples covering a range as wide as possible, taking into account seasonal, geographical and genetic variations. The number of samples for the adjustment should be sufficient for the statistics used to check the performance. For the determination of the bias, at lea
47、st 10 samples are needed, for the determination of standard error of prediction (SEP) and for the slope adjustment, at least 20 samples are needed. 7.5 Checking instrument stability See EN ISO 12099:2010, Clause 9. 7.6 Follow up of method performance Performance of the method shall be checked at lea
48、st annually, against reference methods to secure the constant adequacy of the model with the requirements of this standard (see 5.4.2). This performance test shall be made on samples selected from the pool of analyzed samples. It may be necessary to resort to some sampling strategy to ensure a balan
49、ced sample distribution over the entire calibration range and to ensure that samples with a commercially important range are covered. At least 20 samples are needed (to expect a normal distribution of variance). For instruments operated in a network and adjusted against a master instrument, it is sufficient to run the performance check of the method of this last one. The adjustment (7.4) respecting the requirements of this clause may be used for the follow-up of the method performance. It is recommended to participate in an internationally ac
