Exploring the Possibility to Shorten the Duration and Reduce the Number of Replicates in Biomethane Potential Tests (BMP)

Biochemical methane potential (BMP) tests are the most reliable method for the direct evaluation of the methane yield from a specific feedstock in anaerobic digestion. However, these tests are time-consuming (about 1 month) and quite expensive (need of no less than two or three replicates). This study evaluates the accuracy of the "first-order kinetic", "logistic" and "Gompertz" models in predicting the BMP values, calibrating the models' parameters with the data collected in shorter BMP tests (i.e., 5, 7, 10, 14 and 21 days) than usually (28 days or more). Moreover, the influence of the number of replicates (i.e., two or three) on the model prediction accuracy was also evaluated. A database from 32 BMP tests, previously carried out on different substrates, was adopted for these evaluations. The test duration significantly influences the prediction accuracy for two models (Gompertz and first-order kinetic), while the number of replicates is less influencing. The ultimate methane production is not accurate if the models use parameters from short (less than 10 days) BMP tests. The increase in test duration to 21 days gives BMP predictions with errors below 10% for Gompertz and logistic models.[GRAPHICS].