https://www.euro-online.org/enog/inoc2007/Papers/mac-slots.html https://www.euro-online.org/enog/inoc2007/Papers/m https://www.euro-online.org/enog/inoc2007/Papers/mac-slots.html

NIRs Model Validation

sabki's picture
Forums: 

Dear All,

I have used NIRs FOSS (XDS-RCA), last year I did calibration model for wood properties especially screen pulp yield (SPY) collaborated with Laboratory X in other country. When I send samples to Lab X for Lab analyses we also do analyses in my Lab. (correlation Lab X and my lab :0.85). In NIRA Model for reference data I still used from Lab X data. Could I do validation after calibration only in my laboratory or I must send samples to Lab X?

Regards,

Sabki

shileyda's picture

Sabki, the answer is yes to both. You should continue to validate to Lab X, but at the same time you should also measure at your own lab to develop an understanding of how your lab compares to Lab X. Once you have compared to Lab X you could change to your own lab if the relationship was good enough for you.  We have worked extensively with pulp yield and found that different labs had extreme variation in between their results, but also there can be significant variation even at a single lab. Your quality control program should also include blind duplicates of the same sample at the labs to determine the standard error of laboratory (SEL).

Regards,

Dan