[Feasibility of the extended application of near infrared universal quantitative models]

Construction of a successful near infrared analysis model is a complex task. It spends a lot of manpower and material resources, and is restricted by sample collection and model optimization. So it is important to study on the extended application of the existing near infrared (NIR) models. In this paper, cephradine capsules universal quantitative model was used as an example to study on the feasibility of its extended application. Slope/bias correction and piecewise direct standardization correction methods were used to make the universal model to fit to predict the intermediates in manufacturing processes of cephradine capsules, such as the content of powder blend or granules. The results showed that the corrected NIR universal quantitative model can be used for process control although the results of the model correction by slope/bias or piecewise direct standardization were not as good as that of model updating. And it also indicated that the model corrected by slope/bias is better than that by piecewise direct standardization. Model correction provided a new application for NIR universal models in process control.