Abstract
Biogas production through anaerobic digestion (AD) plays an important role in the effort to reduce the amount of greenhouse gas emissions from energy production. At the same time commercial digestion processes are commonly operated well below their maximal capacity. One reason for this is the lack of suitable sensors for monitoring key parameters in the process. Near infrared (NIR) spectroscopy is seen as a potentially interesting technique for liquid phase measurements in AD processes. In this work, the concentrations of the important constituents, ammonium, acetate, propionate and total volatile fatty acids (TVFA), were measured by combining NIR transmittance measurements and partial least-squares (PLS) models. In order to obtain additional calibration data without having to perform additional reference measurements, and in order to enhance the constituent specific correlation in the calibration data, a spiking procedure was implemented according to a central composite design. This methodology reduced the root mean square error of prediction (RMSEP) for ammonium from 176 mg L−1 to 127 mg L−1, for acetate from 334 mg L−1 to 260 mg L−1 for propionate from 258 mg L−1 to 203 mg L−1 and for TVFA from 858 mg L−1 to 704 mg L−1. The spiking procedure also significantly increased the correlation between model predictions on pure constituent spectra and reference values, compared to that of models based on non-spiked data.
© 2013 IM Publications LLP
PDF Article
More Like This
Cited By
You do not have subscription access to this journal. Cited by links are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.
Contact your librarian or system administrator
or
Login to access Optica Member Subscription
Add to BibSonomy