Real-time monitoring of powder blend composition using near infrared spectroscopy

Abstract

Near Infrared Spectroscopy (NIRS) is a very powerful utility in a Process Analytical Technology (PAT) system because it can be used to monitor a multitude of process parameters non-invasively, non-destructively in real time and in hazardous environments. A catch to the versatility of NIRS is the requirement for Multi-Variate Data Analysis (MVDA) to calibrate the measurement of the parameter of interest. This paper presents a NIRS based real time continuous monitoring of powder blend composition which has widespread applications such as the pharmaceutical industry. The proposed system design enables reduction of optical path length so that the sensors can be successfully installed into powder conveyance systems. Sensor signal processing techniques were developed in this work to improve accuracy while minimizing pre-processing steps. The paper presents the implementation of several parameter estimation methodologies applied to sensor data collected using MATLAB® software for a model powder blending process. Several techniques were examined for the development of chemometric models of the multi-sensor data, including Principal Component Analysis (PCA), Partial Least Squares Regression (PLSR), Support Vector Machines (SVM) and Artificial Neural Networks (ANN). The performances of each of the models were compared in terms of accuracy (MSE) in predicting blend composition. The results obtained show that machine learning-based approaches produce process models of similar accuracy and robustness compared to models developed by PLSR while requiring minimal pre-processing and also being more adaptable to new data.