Acoustic chemometric prediction of total solids in bioslurry: A full-scale feasibility study for on-line biogas process monitoring

Dry matter is an important process control parameter in the bioconversion application field. Acoustic chemometrics, as a Process Analytical Technology (PAT) modality for quantitative characterisation of dry matter in complex bioslurry systems (biogas fermentation), has not been successful despite several earlier dedicated attempts. A full-scale feasibility study based on standard addition experiments involving natural plant biomass was conducted using multivariate calibration (Partial Least Squares Regression, PLS-R) of acoustic signatures against dry matter content (total solids, TS). Prediction performance of the optimised process implementation was evaluated using independent test set validation, with estimates of accuracy (slope of predicted vs. reference values) and precision (squared correlation coefficient, r2) of 0.94 and 0.97 respectively, with RMSEP of 0.32% w/w (RMSEPrel = 3.86%) in the range of 5.8–10.8% w/w dry matter. Based on these excellent prediction performance measures, it is concluded that acoustic chemometrics has come of age as a full grown PAT approach for on-line monitoring of dry matter (TS) in complex bioslurry, with a promising application potential in other biomass processing industries as well.

► Acoustic spectra of various quantities of total solids in bioslurry were acquired.
► Partial Least Squares Regression method was used to predict the total solids.
► Test-set validation was applied for model assessment.
► Good accuracy and precision were obtained.
► Total solid range of 5.8 to 10.8 weight percent dry matter was predicted.