Early warning of disturbances in a laboratory-scale MSW biogas process

The use of near-infrared spectroscopy (NIR) to monitorthe dynamics of a biogas process was evaluated using multivariate data analysis. The digester was a completely stirred 8 I tank reactor fed with the organic fraction of source-sorted MSW. Intermittently the digester was overloaded with feed. Before and after overload on-line monitoring of NIR spectra and off-line analysis in the liquid and the gas phase of traditional chemical variables and microbial biomass, determined as total concentration of phospholipid fatty acids (PLFA and PLEL), were done. The dynamics that occurred due to overloading could be followed using principal component analysis of the obtained NIR-spectra. In addition, the response to changes in the digester fluid was reproducible and could be detected within five minutes, which can be considered as real-time monitoring. Selected wavelengths in the region 800-2,000 nm were used to make a PLS1 -regression with propionate. The regression resulted in a good correlation for propionate (R= 0.94 and RMSEP of 0.21 g/l in the range of 0.3-3 g/l). The results indicate the possibility to develop an early warning biogas control system based on near-infrared spectroscopy monitoring of propionate.