ISO 23291:2020 pdf free download.Milk and milk products — Guidelines for the application of in-line and on-line infrared spectrometry.
1 Scope
ISO 23291v gives guidelines For using infrared spectrometry in in-line and on-line applications for dairy processing, These applications are distinct to those covered in Iso 21543 I IDF 201.
It Is applicable, but not limited, to:
— the determination of protein, fat and total solids in liquid milk and milk products using mid and near infrared spectrometry;
— the determination of protein, tat and moisture in solid or semi-solid products, such as milk powder, and butter and liquid dairy streams using near infrared spectrometry.
2 NormatIve references
There are no normative references in ISO 23291.
3 Terms and definitions
For the purposes of ISO 23291, the following terms and definitions apply.
ISO and IEC maintarn terininological databases for use In standardization at the following addresses:
— ISO Online browsing platform: available at https:/Jwwwiso.orEjobp
— IEC Electropedia: available at httpilwww.electropedia.orRJ
3.1
in-line analysis
analysis of a product line where the sensor probe interfaces directly with the product stream being measured, or a reflectance measurement through an optical window into the product stream
3.2
on-line analysis
analysis of a product line where the sensor probe interfaces indirectly with the product stream being measured by way of a bleed loop, automated grab sampler or other means of subsampling
3.3
at-line analysis
analysis of a product where the instrument is physically remote to the product stream being measured and the sample is manually introduced to the instrument
Note Ito entry: While not covered in this document, this definition is added here in order to dis Unguish this type of specromefric analysis From In- and on-line apparatus.
3.4
near infrared instrument NIR instrument
proprietary apparatus utilizing wavelengths within the range 400 nm to 2 500 nni or 25 000 cm1 to 4 000 cm1 (both visible and NIR range) or 12 1120 cm1 to 4 000 cm1 (NIR range only) which, when used under certain conditions, estimates mass fractions or other parameters of use
3-5
mid infrared Instrument MIR instrument
proprietary apparatus utilizing wavelengths in the range 4 000 cm1 to 400 cmi, which, when used under the conditions specified In this document, estimates the mass fractions or other parameters of use specified in Cbuse I
4 Principle
A laboratory in-hne or on-line instrument is installed according to the manufacturer’s guidelines [or the type of process under measurement. Absorbance within the wavelength regions mentioned above is measured by transmission, reflectance and a combination of both, or by attenuated total reflectance (ATR), The resulting spectral Information Is transformed to constituent concentrations or constituent values with other units by calibration models developed by representative samples from the population to be tested.
5 Apparatus
5.1 Infrared Ilistruinent. based on diffuse reflectance or transmittance measurement In the near (400 nm to 2 500 nm or 25 000 cm to 4 000 cm1 or 12 820 cm to 4 000 cm-’) or mid (4 000 cm-1 to 400 cm 1) wavelength region or segments of this or at discrete wavelengths. The optical operation principle may be dispersive (cij, grating monachromators). interfcrometrlc or diode-array based. The instrument should be provided with an appropriate diagnostic test system for testing photometric instrument noise, wavelength accuracy and wavelength precision (for scanning spectrophotometers). The instrument shall be able to optically view Into the product stream with an appropriate interface. There are many commercial such devices with a wide variety of technology depending on specific applications.
6 Installation and sampling considerations
6.1 General
Any installation shall cover both the integrity of the infrared instrument as well as inter’ace with the
process stream, Key aspects for the integrity of the instrument include:
— protection from cleaning regimes;
— isolation from vibration, dust and other environmental contaminants;
— an appropriate temperature regime for the specific instrument;
— many regions have strict protocols [or materials in contact with human food products, such as Regulation (EC) No 1935j2004ll or US equivalent 3A, and this Is an important consideration for probes or cell construction.
A reliable, stable and consistent sampling interface is the key to successful use of in-line and on-line spectrometry. The following aspects are important.
— The ability to sample a representative flow of product. This can be ascertained by experimentation as well as an understanding of the fluid dynamics of the process [low.
— The ability to be consistently cleaned to the same level as the rest of the installation for good grade hygiene. For liquid product streams, this may mean that the probe is cleaned by the regular cleaning an place protocols. For powder, an air frt or similar can be necessary to remove sample prior to each measurement, Experimentation is often required to determine the most effective cleaning protocol [or a specific environment.
8 CalibratIon and validation
It is noted that calibration and validation is a complex area and many aspects are common to both at- line as well as in-line or online Instruments, These are well covered In Iso 21543 I IDE 201.
The instrument shall be calibrated before it is used. Because of the complex nature of the infrared wavelength regionIl, the Instrument shall be calibrated on a series of process samples with corroborating laboratory reference tests or similar analytical results. It is noted that when the process is statistically under control, recognized secondary methods such as at-line NIR/Fourier-transform infrared spectroscopy (FTIR) can also be used to develop and periodically check the performance of calibration models. However, it is strongly recommended that the bulk olthe spectra used for calibration are measured under in-line or on-line process conditions. An In-line Instrument that is located at-line or In a laboratory may be used with care to extend the range of calibration samples, but these should not comprise the bulk of the calibration set. This will ensure spectral variation inherent to the process is adequately captured in the calibration spectra set. Due to the nature of NIR spectra, typically an NIR instrument will require a greater number of calibration samples than a MIR instrument, but this can vary depending on the sample matrix type and the specific compositional parameter being calibrated. If a secondary instrument is used as reference for the calibration and validation work, then it is essential to ensure it is In adequate control. When using chemical analysis as reference analysis, it Is highly recommended to use duplicate analyses instead of a single analysis.
If the required agreement and precision cannot be obtained by a single calibration, then the application area can be split up into static or dynamic sub-areas, each with an associated calibration, to fulfil the requirements. These would ideally be selected automatically when required in conjunction with changes in processing conditions. A nonlinear modelling approach could also be used. Here, an identification step could be used to know which sample/product is currently to be measured. A model selection by a trigger from the process control system can also be a useful adjunct to ensure the appropriate models are in use for different products.
With in-line instruments, there is often a challenge 10 obtain sufficient variation In sample composition. Ideally, deliberate process variation can be used to obtain an appropriate sample range. If this is not possible, samples taken from start up or during processing changes can yield greater variation, provided care is taken that they are truly representative of the process.
Calibration models developed for in-line instruments should have good performance in terms of agreement and precision as described in ISO 21543 I IUF 201. In some cases, where the reaction time of the process is much longer than the in-line instrument measurement time, a plot of moving average of results can be very helpful to increase the perFormance of the instrument reading. When averaging can increase accuracy of the instrument reading or trending, the results can be used in a better way to control the process. However, even more important is to have good performance with respect to the repeatability In order to detect small variations in the production. It Is also possibLe to use in-line infrared measurement for trend control. In cases of a biased reading, the results can still be used to control the process based on trending information.
Validation should be accomplished by checking the performance of the installation and associated calibration using an independent test set, preferably sampled after the calibration period but also covering a similar amount of process variation, If possible. This data can be analysed as described In ISO 21543 I IDE 201 and the calibration adjusted if required. Following calibration adpustment, revalidation should be carried out. A validation report specifying the following information should also be prepared:
— all the information necessary for complete identification of the samples (e.g. date, sample type):
— the laboratory test method(s) used, with reference to the relevant International Standards and Including a laboratory reference error;
— any unusual circumstances that could have influenced the results;
— the test results obtained.