engleski

Development of NIR-based ANN models for on-line monitoring of glycerol concentration during biodiesel production in a microreactor

During the production process, a whole range of analytical methods must be developed to monitor the quality of production and the desired product(s). Most of those methods belong to the group of off-line monitoring methods and are usually recognized as costly and long-term. In contrast, on-line monitoring methods are being used as fast, reliable, simple, and repeatable. The main objective of this study was to compare different methods for monitoring total glycerol concentration as one of the indicators of process efficiency during biodiesel production in a batch reactor and in a microreactor. During the biodiesel production process, glycerol concentration was measured off-line using standard methods based on UV-VIS spectrophotometry and gas chromatography. Neither method provided satisfactory results, namely, both analyses showed significant deviations from the theoretical value of glycerol concentration. Therefore, near infrared spectroscopy (NIR) analysis was performed as an alternative analytical method. The analysis using NIR spectroscopy was performed in two ways: off-line, using a sample collected during the transesterification process, and on-line by continuous measurement of glycerol concentration in a rector. Obtained results showed a great NIR application potential not only for off-line but also for on-line monitoring of the biodiesel production process.

biodiesel ; glycerol ; near infrared spectroscopy ; artificial neural networks ; on-line measurements

nije evidentirano