In a pilot plant study, 36 batches of margarine were produced, with a moisture range of 14.28–21.91 weight-%. The margarine was measured in- and off-line using a near infrared (NIR) instrument and an off-line laboratory reference method. For the in-line measurements, diffuse transmittance fibre-optic probes were used. For the off-line measurements, sample cups were used. Partial least square regression of multiplicative scatter corrected NIR spectra, in the 780–1100 nm region, gave prediction errors, expressed as root mean square error of cross-validation ( RMSECV) of 0.13 weight-% moisture for the in-line measurements. The corresponding validation correlation coefficient ( R) was 0.998. The corresponding RMSECV and R values for the off-line measurements were 0.14 and 0.997, respectively. The prediction error results are considered to be very good. It can be concluded that in- and off-line NIR determination of moisture in margarine gave very good prediction results and could be used in production.
No takes yet. Share an insight, caveat, or question.
Isaksson et al. (2001) studied this question.
Synapse has enriched one closely related paper. Consider it for comparative context: